Sinkron : jurnal dan penelitian teknik informatika

Paraphrase Generation For Reading Comprehension

Faishal Januarahman, Ade Romadhony — Sun, 01 Oct 2023 00:00:00 +0700

Reading comprehension is an assessment that tests readers understanding of a concept from the given text. The testing process is conducted by providing questions related to the content within the context of the text. The purpose of this research is to create new question variations from existing questions, and one of the methods to achieve this is by paraphrasing questions through the task of paraphrase generation. This can help ensure that readers have fully grasped a concept of a text. This study employs a traditional approach known as the thesaurus-based approach, in which the process involves substituting synonyms using the Indonesian Thesaurus dictionary. The data used consists of a list of Indonesian language reading comprehension assessment questions ranging from elementary to high school levels. To measure the quality of the generated paraphrased questions, two evaluation processes are conducted which are automatic evaluation with the scores ranging from 0-1 and human evaluation with score ranging from 1-4. The automatic evaluation includes the BLEU-4 metric, resulting in a score of 0.044, and the ROUGE-L metric, resulting an F1-score of 0.421. As for human evaluation, the obtained relevancy score is 2.533, and the fluency score is 3.186. The results from both evaluation metrics indicate that the generated paraphrased questions exhibit diverse new word choices but tend to have slightly different meanings compared to the reference questions.

Comparison Analysis of C4.5 Algorithm and KNN Algorithm for Predicting Data of Non-Active Students at Prima Indonesia University

Sun, 01 Oct 2023 00:00:00 +0700

Education is important nowadays because universities need to improve their students' skills so they can compete in the globalization era. Education can be obtained through both formal and informal channels, and knowledge is available everywhere, especially in today's world where information tools are rapidly evolving. Inactive students are students who do not participate in a course for a maximum of two consecutive semesters. Students who are not active have the opportunity to drop out of university studies. Students who drop out of college are usually motivated by economic factors, and the cessation of the lecture process can cause inactivity and administrative costs. Therefore, this research was conducted using the C4.5 algorithm method and the K-Nearest Neighbor (KNN) algorithm to compare and predict data on inactive students at Universitas Prima Indonesia. The research continued with the data collection and data preprocessing stages, after which the data mining process was carried out to get the final results of this research. The testing process follows the process of comparing the C4.5 algorithm and the K-Nearest Neighbor (KNN) algorithm with K-fold crossing. This evaluation step is compared by considering the comparison values of the confusion matrix (precision, precision, recall). The accuracy results obtained by each algorithm provide information about the effectiveness of using these techniques in processing the specified dataset. The accuracy of the Decision Tree C4.5 algorithm is 99.12% and the K-Nearest Neighbors algorithm is 99.14%. Based on research conducted using the K-Nearest Neighbors and C4.5 algorithms to predict inactive students, the KNN algorithm is more accurate than the C4.5 algorithm.

Decision Support for Selection of The Best Teachers Recommendations MCDM-AHP and ARAS Collaborative Methods

Akmaludin, Erene Gernaria S. , Rinawati , Ester Arisawati , Linda Sari Dewi — Sun, 01 Oct 2023 00:00:00 +0700

The role of the teacher is very important for the progress of the nation which can increase the dignity of the nation. The quality of education can increase thanks to the support of teachers who have good dedication in developing the learning process, especially in curriculum development. The teacher's biggest contribution is to make students ready to become the nation's credible successors. The purpose of this study is to provide an assessment in the selection process of teachers in an objective and selective manner. Method recommendations that can be raised in this study are collaborative methods that play a role in multi-criteria selection, namely MCDM-AHP and ARAS. Both of these methods can be said to be able to implement a selective and credible selection process for teachers which is carried out through the stages of data conversion and normalization which in turn can determine decision support in the multi-criteria selection process. The results of this study provide the best solution in selecting alternatives with a multi-criteria barometer for measurements with the ARAS method. The resulting decision support for the selection process of twenty teachers resulting in the best assessment can be seen from the optimization function which is based on the maximum value default divisor basis. The results obtained from obtaining the highest utility value of the twenty alternatives, the top three values that can be drawn as support for decision making are owned by A3 with a utility weight of 0.891, followed by the two highest ratings A17 and A14 with respective weights of 0.888 and 0.884.

Ontology-based Nutrition Recommender System for Stunting Patients

Nur Laili Ramadhani, Z. K. A. Baizal — Sun, 01 Oct 2023 00:00:00 +0700

Stunting is a growth disorder that occurs in early childhood. This condition occurs because the child has a chronic nutritional problem which triggers the child to have a height below normal. The indicator used as a standard for whether a child is stunted or not is height for age. If a child has a z-score value less than -2 standard deviations, then the child is said to suffer from stunting. Poor nutritional intake is one of the factors causing children to suffer from stunting. Most Indonesian people think that the genetics of both parents causes children to be shorter than their age, but genetics is a minimal factor that causes stunting. In 2020, Indonesia ranks second in the prevalence of stunting in Southeast Asia, according to the Asian Development Bank (ADB) report. Based on the results of the Indonesian Nutritional Status Survey (SSGI) in 2021, the stunting prevalence rate in Indonesia 2021 is 24.4%, but in 2022, the stunting prevalence rate will drop to 21.6%. One way to treat stunting in children is by providing daily nutritional intake according to the child's condition. In this study, we used the Telegram chatbot with an ontology and the rules Semantic Web Rule Language as a knowledge base. The accuracy performance of our system is 93.3% which shows that our system can provide nutritional recommendations for stunting patients.

Design of a Home Door Security System Based on NodeMCU ESP32 Using a Magnetic Reed Switch Sensor and Telegram Bot Application

Syahri Ramadhani, Dhanny Permatasari Putri — Sun, 01 Oct 2023 00:00:00 +0700

With very rapid technological advancements, it is not possible now that all activities can be carried out quickly, easily, and instantly. The Internet of Things (IoT) allows us to solve various problems by making some devices communicate with each other across the virtual world network. This study uses the prototyping method by explaining how to design a home door security system that can be controlled by a smartphone with a WiFi connection and can use a card consisting of a chip connected to a reader. The design of this tool consists of several stages, namely designing a block diagram of how the circuit works, and designing hardware (Hardware) and software (Software). From the results of system testing, it can be seen that when conducting RFID testing by tapping the card on the reader. The system is successful and the door can be opened according to the NUID card that has been registered with the program and it can fail by using a card that is not registered in the system. Which indicates the test results are working properly. The working capability of the system on the door security device is as expected and the response from the magnetic reed switch sensor as input from the notification is very good. This system using RFID functions to make it easier for users to control the door of the house with a smartphone connected to wifi so that users don't just use conventional keys.

Search Optimization of PIP Scholarship Recipients In Web-Based Student Data Application Using The Levenshtein Distance Algorithm

Yoga Handoko Agustin, Yosep Septiana, Arbi Yuan Aspahany — Sun, 01 Oct 2023 00:00:00 +0700

Realizing that education is very important, the government supports every citizen to get education. One of the government programs is the Smart Indonesia Program. PlP is a scholarship designed to help school-age children from poor/vulnerable families to continue to receive education services, both through formal elementary to high school/vocational schools and non-formal pathways from package a to package c and special education. SDN II Babakanloa has not been touched by technology for processing student data. So that the student section has difficulties in recording and updating student data. Student names have unique identities and errors often occur in typing the keywords to be searched. This results in an information that is desired or sought can not be found. Therefore we need a web-based data application that can provide keyword corrections in searching for student names. This study aims to create a web-based student data application by optimizing corrections to typing keywords searched by implementing the Levenshtein Distance Algorithm and also making it easier to process and search student data. The development method used is the Rational Unified Process (RUP) with the stages of Inception, Elaboration, Construction, and Transition. Designed using the CodeIgniter Framework with the PHP and JavaScript programming languages. The application of the Levenshtein Distance Algorithm can optimize the search for student data and reduce the occurrence of search errors by School Operators. The application of the Levenshtein Distance Algorithm produces a very good accuracy rate of 94% of the results of student data correction. accordance with the expectations of the School Operator. So it shows that the application of the Levenshtein Distance Algorithm is appropriate to use in optimizing the search.

The Performance of the Equal-Width and Equal-Frequency Discretization Methods on Data Features in Classification Process

Sun, 01 Oct 2023 00:00:00 +0700

The classification process often needs help with suboptimal accuracy values, which can be attributed to various factors, including the dataset's wide range of attribute values. Discretization methods offer a solution to address these issues. This study aims to compare the effectiveness of Equal-Width and Equal-Frequency discretization methods in enhancing accuracy during the classification process using datasets with varying sizes. The research employs Naïve Bayes, Decision Tree, and Support Vector Machine as classification models, with three datasets utilized: Bandung City Traffic data (3804 records), Bandung City COVID-19 cases data (2718 records), and Bandung City Dengue Fever Disease Index data (150 records). Three experimental scenarios are executed to assess the impact of the two discretization methods on accuracy. The first scenario involves no discretization, the second employs Equal-Width, and the third applies Equal-Frequency discretization. Experimental results indicate significant accuracy improvements post-discretization. The Naïve Bayes model achieved 94% accuracy for the Traffic dataset, while the Decision Tree achieved 71% accuracy for the COVID-19 dataset and an impressive 98% for the Dengue Fever Disease dataset. These outcomes demonstrate that applying Equal-Width and Equal-Frequency discretization methods addresses the challenge of wide attribute value ranges in the classification process.

Comparison of Algorithms for Sentiment Analysis of Operator Satisfaction Level for Increasing Neo Feeder Applications in PDDikti Higher Education LLDIKTI Region VI Semarang Central Java

M. Ulil Albab, Ema Utami, Dhani Ariatmanto — Sun, 01 Oct 2023 00:00:00 +0700

Sentiment analysis on the satisfaction level of PDDikti operators is very important to find out how PDDikti operators feel after the version of the academic reporting application for higher education was upgraded, namely Neo Feeeder. The increase in the version of this application causes some of the features in it to not function properly. So some academic reporting activities from tertiary institutions experience problems. As a result of this condition, the most felt impact is students, where students experience delays in graduation. Then it is necessary to evaluate through sentiment analysis from PDDikti operators to find out the response from operators and be able to provide positive suggestions to developers from the PDDikti reporting application. This study applies several classification methods for sentiment analysis at once, including the Random Forest algorithm, the Support Vector Machine algorithm, the Multinomial Naïve Bayes algorithm, the Decision Tree algorithm, and the K-Nearest Neighbor algorithm. Of the 5 methods applied, the results of their performance accuracy will be compared. The performance of the highest classification algorithm is the K-Nearest Neighbor (K-NN) algorithm which produces an accuracy value when testing data, which is up to 90% using the oversampling technique in unbalanced classes. While the lowest classification accuracy performance value is in the Multinomial Naïve Bayes (MNB) algorithm with a value of 76%. It is proven that oversampling can help the performance of the classification algorithm to be more optimal. Thus, it should be noted that the balance of data classes is an important factor when applying the classification method.

Two-Stage Sentiment Analysis on Indonesian Online News Using Lexicon-Based

Vinardo, Ito Wasito — Sun, 01 Oct 2023 00:00:00 +0700

The image of a supplier company is often associated with the well-known brand it supplies, and its reputation can be influenced by online news circulation. To maintain a positive image, it is crucial for the company to monitor and manage online news to rectify any false information. Failure to maintain a good company image can lead to customer order loss and even company shutdown.

This paper aims to conduct a two-stage sentiment analysis on Indonesian news articles regarding unilateral layoffs by company XYZ. The first stage will analyze sentiment in the circulating news about the layoffs, while the second stage will assess sentiment after the company releases a press release to provide accurate information. The VADER lexicon-based method, utilizing the InSet and SentiStrength_ID Indonesian dictionaries, will be employed to analyze sentiment before and after the press release. This will enable us to compare sentiment and evaluate the effectiveness of the press release and the Indonesian dictionaries in analyzing sentiment in the news. The research findings indicate that the company's press release, aimed at correcting false information, had a positive impact by reducing negative sentiment and generating a more positive sentiment in the second stage. Moreover, the selection of the sentiment analysis dictionary also plays a critical role in determining the sentiment analysis results.

Electronic Product Recommendation System Using the Cosine Similarity Algorithm and VGG-16

Irfan Rasyid, Muhammad Resa Arif Yudianto, Maimunah, Tuessi Ari Purnomo — Sun, 01 Oct 2023 00:00:00 +0700

The recommendation system is a mechanism for filtering a batch of data into numerous data sets based on what the user wants. Cosine similarity is one of the algorithms used in creating recommendation model. This algorithm employs a calculation approach between two things by measuring the cosine between the two objects to be compared. Image-based recommendation systems were recently introduced since word processing to generate recommendations had the issue of duplicating product descriptions for different types of items. Before processing with cosine similarity, image feature extraction requires the use of a deep learning algorithm, VGG16. The purpose of this research is to make it easier for customers to select the desired electronic goods by providing product recommendations based on product visual similarity. This model is able to recommend 10 products that are similar to the selected product. The presented product has a cosine value near one, and the discrepancy with the selected product's cosine value is modest. The mAP technique was used for model testing, and the smartwatch category received the greatest mAP value of 94.38%, while the headphone category had the lowest value of 70.84%. The average mAP attained is 81.50%. These findings show that mAP accuracy varies by category. This disparity is due to the unequal dataset in each category.

Classification of E-Commerce Product Descriptions with The Tf-Idf and Svm Methods

Dagobert Pakpahan, Veronika Siallagan, Saut Siregar — Sun, 01 Oct 2023 00:00:00 +0700

The rapidly growing e-commerce sector presents a significant challenge in navigating an abundance of products. Understanding and classifying product descriptions efficiently and accurately is crucial to improving user experience and business operations. This research employed the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm and Support Vector Machine (SVM) for the classification of e-commerce product descriptions into four categories: Electronics, Household Items, Books, and Clothing. The initial phase involved pre-processing of text data which incorporated text cleaning, tokenization, part-of-speech tagging, entity recognition, and conversion into a vector representation. The resulting model was trained and tested using the SVM algorithm. Our model demonstrated a high degree of accuracy, achieving 99.2% during the training phase and 95.7% in the testing phase. This model provides a valuable tool for e-commerce businesses, as it allows for accurate classification of products based on their descriptions. This could lead to improved user navigation and overall user experience on e-commerce platforms.

Breast Cancer Detection in Histopathology Images using ResNet101 Architecture

Maie Istighosah, Andi Sunyoto, Tonny Hidayat — Sun, 01 Oct 2023 00:00:00 +0700

Cancer is a significant challenge in many fields, especially health and medicine. Breast cancer is among the most common and frequent cancers in women worldwide. Early detection of cancer is the main step for early treatment and increasing the chances of patient survival. As the convolutional neural network method has grown in popularity, breast cancer can be easily identified without the help of experts. Using BreaKHis histopathology data, this project will assess the efficacy of the CNN architecture ResNet101 for breast cancer image classification. The dataset is divided into two classes, namely 1146 malignant and 547 benign. The treatment of data preprocessing is considered. The implementation of data augmentation in the benign class to obtain data balance between the two classes and prevent overfitting. The BreaKHis dataset has noise and uneven color distribution. Approaches such as bilateral filtering, image enhancement, and color normalization were chosen to enhance image quality. Adding flatten, dense, and dropout layers to the ResNet101 architecture is applied to improve the model performance. Parameters were modified during the training stage to achieve optimal model performance. The Adam optimizer was used with a learning rate 0.0001 and a batch size of 32. Furthermore, the model was trained for 100 epochs. The accuracy, precision, recall, and f1-score results are 98.7%, 98.73%, 98.7%, and 98.7%, respectively. According to the results, the proposed ResNet101 model outperforms the standard technique as well as other architectures.

Vehicle Detection and Identification Using Computer Vision Technology with the Utilization of the YOLOv8 Deep Learning Method

Sun, 01 Oct 2023 00:00:00 +0700

Vehicle identification and detection is an important part of building intelligent transportation. Various methods have been proposed in this field, but recently the YOLOv8 model has been proven to be one of the most accurate methods applied in various fields. In this study, we propose a YOLOv8 model approach for the identification and detection of 9 vehicle classes in a reprocessed image data set. The steps are carried out by adding labels to the dataset which consists of 2,042 image data for training, 204 validation images and 612 test data. From the results of the training, it produces an accuracy value of 77% with the setting of epoch = 100, batch = 8 and image size of 640. For testing, the YOLOv8 model can detect the type of vehicle on video assets recorded by vehicle activity at intersections with. However, the occlusion problem overlapping vehicle objects has a significant impact on the accuracy value, so it needs to be improved. In addition, the addition of image datasets and data augmentation processes need to be considered in the future

Automatic OEE Data Collection and Alert System for Food Industry

Ruly Sumargo, Amelia Makmur — Sun, 01 Oct 2023 00:00:00 +0700

The constant demand for food and beverages to sustain human life drives fierce competition among manufacturers, focusing on product excellence in terms of timeliness, quality, and pricing. The key to competitiveness depends in optimizing manufacturing processes by efficiently utilizing company resources. To ensure the overall optimization and reliable flow of manufacturing processes, a systematic evaluation process must be used, Overall Equipment Efficiency (OEE) stands out as a prominent performance measurement metric in manufacturing process efficiency. OEE serves as a valuable diagnostic tool, exposing areas for improvement and losses transparently. Accurate OEE measurement necessitates the implementation of an automated data collection system with minimum human dependencies, human intervention, and conducting on-the-fly calculations to informed the stakeholder/user. Data quality and accuracy in OEE measurement is very critical. Low quality and accuracy data could lead to false decision. OEE categorizes losses into six groups loss to pinpoint significant factors for potential improvement. Once OEE could be maintain at high level with high data accuracy and right improvement point, an optimum manufacturing process, and cost effective in manufacturing expenses will be achieve. Base on the result comparison for OEE result before and after the system implementation, positive improvement in OEE could reach 8.06%. This scenario be adopted by other company, and could become a model for 1^st phase journey in company digital transformation.

Comparative Analysis of CNN and CNN-SVM Methods For Classification Types of Human Skin Disease

Dendi Anggriandi, Ema Utami, Dhani Ariatmanto — Sun, 01 Oct 2023 00:00:00 +0700

Cancer is one of the leading causes of death worldwide, with skin cancer ranking fifth. The skin, as the outermost organ of the body, is susceptible to various diseases, and accurate diagnosis is crucial for effective treatment. However, limited access to dermatologists and expensive skin biopsies poses challenges in achieving efficient diagnosis. Therefore, it is important to develop a system that can assist in efficiently classifying skin diseases to overcome these limitations. In the field of skin disease classification, Machine Learning and Deep Learning methods, especially Convolutional Neural Network (CNN), have demonstrated high accuracy in medical image classification. CNN's advantage lies in its ability to automatically and deeply extract features from skin images. The combination of CNN and Support Vector Machine (SVM) offers an interesting approach, with CNN used for feature extraction and SVM as the classification algorithm. This research compares two classification methods: CNN with MobileNet architecture and CNN-SVM with various kernel types to classify human skin diseases. The dataset consists of seven classes of skin diseases with a total of 21.000 images. The results of the CNN classification show an accuracy of 93.47%, with high precision, recall, and F1-score, at 93.55%, 93.74%, and 93.62%, respectively. Meanwhile, the CNN-SVM model with "poly," "rbf," "linear," and "sigmoid" kernels exhibits varied performances. Overall, the CNN-SVM model performs lower than the CNN model. The findings offer insights for medical image analysis and skin disease classification research. Researchers can enhance CNN-SVM model performance with varied kernel types and techniques for complex feature representations.

Prediction of the Human Development Index for Equitable Development in West Sumatra Province Using the C4.5 Algorithm

Weri Sirait, Nur Azizah — Sun, 01 Oct 2023 00:00:00 +0700

Unequal development in Indonesia can be seen from the Human Development Index. The Human Development Index is a tool used to measure the attainment of the quality of life of a region or country and as material for economic policy on quality of life. It contains components of health level, education level and welfare level. In 2022, West Sumatra Province achieved the 9th highest Human Development Index in Indonesia, namely 73.26, with this figure the West Sumatra Province Human Development Index is above the national average. However, there are still regencies/cities in West Sumatra Province that have achievements below the national average. This factor causes the development conditions in West Sumatra Province to be uneven. Uneven human development conditions will make it difficult for the government to improve Human Resources (HR). In this research, the C45 Data Mining Algorithm was implemented to predict the Regency/City Human Development Index in West Sumatra Province. As is the method of the Central Bureau of Statistics in measuring the Human Development Index, the variables used from the Human Development Index indicators are Life Expectancy, Years of School Expectation, Average Length of Schooling, and Per Capita Expenditures. The Central Statistics Agency data used in this research covers all regencies/cities in West Sumatra during the period 2018-2022. Range levels are grouped into three groups, namely, low, medium, and high. Based on testing using RapidMiner software with the Cross Validation operator, an accuracy value of 86.61% was obtained.

Stock Price Correlation Analysis with Twitter Sentiment Analysis Using The CNN-LSTM Method

Muhammad Noer Ibnu Sina, Erwin Budi Setiawan — Sun, 01 Oct 2023 00:00:00 +0700

The intricate interplay between stock prices, reflecting a company's intrinsic value, and multifaceted factors like economic conditions, corporate performance, and market sentiment, constitutes a vital research domain. Grounded in sentiment analysis, our study deciphers public opinions from vast textual data to gauge sentiment, leveraging Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) models. We focus on Bank Central Asia (BBCA), a prominent Indonesian banking institution, aiming to forecast stock price fluctuations by analyzing sentiment trends extracted from social media, especially Twitter. Meticulous experimentation, encompassing data segmentation, feature extraction, augmentation, and model refinement, yields significant enhancements in prediction accuracy. The CNN-LSTM model's performance improves from 73.41% to a robust 77.75% accuracy, with F1-scores rising from 73.00% to 75.42%. Importantly, strong correlations emerge between sentiment predictions and actual stock price movements, validated by Spearman correlation coefficients. Positive sentiment exhibits a substantial correlation of 0.745 with stock price changes, while negative sentiment exerts notable influence with a correlation coefficient of 0.691. In summary, our study advances the field of sentiment-driven stock price prediction, showcasing deep learning's effectiveness in extracting sentiment from social media narratives. The implications extend to understanding market dynamics and potentially integrating sentiment-aware strategies into financial decision-making. Future research directions could explore model transferability across financial contexts, real-time sentiment data integration, and interpretability techniques for enhanced practicality in sentiment-driven predictions.

Identification of 10 Regional Indonesian Languages Using Machine Learning

Azhar Baihaqi Nugraha, Ade Romadhony — Sun, 01 Oct 2023 00:00:00 +0700

Language Identification plays a pivotal role in deciphering the rich tapestry of Indonesia's diverse regional languages, encompassing a wide spectrum of scripts, and spoken forms. Language Identification, an integral component of Natural Language Processing, is frequently addressed through Text Classification. In this study, we embark on the task of identifying 10 Indonesian languages, leveraging the NusaX dataset, with the overarching objective of contextual language determination. To achieve this, we harness a diverse array of machine learning techniques, including Support Vector Machine, Naïve Bayes Classifier, Decision Tree, Rocchio Classification, Logistic Regression, and Random Forest. We complement these methods with two distinct feature extraction approaches: N-gram and TF-IDF. This comprehensive approach enables us to construct robust models for language identification. Our findings unveil the strong efficacy of these models in discerning Indonesian languages, with the Naïve Bayes Classifier emerging as the frontrunner, achieving an impressive accuracy rate of 99.2% with TF-IDF and an even more remarkable 99.4% with N-Gram. To gain deeper insights, we delve into error analysis, revealing that misclassifications often stem from shared words across different languages. This research is underpinned by the necessity for a robust language identification model, underscoring its critical role within the complex linguistic landscape of Indonesian regional languages. These results hold great promise for applications in automated language processing and understanding within this diverse and multifaceted linguistic context.

Social Media Based Film Recommender System (Twitter) on Disney+ with Hybrid Filtering Using Support Vector Machine

Helmi Sunjaya Ramadhan, Erwin Budi Setiawan — Sun, 01 Oct 2023 00:00:00 +0700

In the current era, the culture of watching TV shows and movies has been made easier by the presence of the internet. Now, watching movies on platforms can be done from anywhere, one of which is Disney+. At times, people find it challenging to decide which film to watch given the multitude of genres and film titles available on these platforms. One solution to this issue is a recommendation system that can suggest films based on ratings. The recommendation system to be utilized involves Collaborative Filtering, Content-Based Filtering, and Hybrid Filtering. This is because Collaborative Filtering and Content-Based Filtering encounter issues like cold start, sparsity, and overspecialization. Thus, the objective of this study is to develop a recommendation system using Hybrid Filtering combined with Support Vector Machine (SVM). In this research, classification will be carried out using poly, linear, and RBF kernels with varying parameters. Techniques such as TF-IDF, RMSE, tuning, and data balancing with SMOTEN will be implemented to enhance accuracy during the classification process. The evaluation employed in this study utilizes the confusion matrix. Support Vector Machine, when tuned and combined with SMOTEN, achieves noteworthy results, particularly with the RBF kernel which attains a Precision score of 0.94. Recall produces a value of 0.93 with the Poly kernel, while the highest Accuracy, at 0.93, is achieved with the RBF kernel. Furthermore, the RBF kernel also demonstrates the highest F1-Score of 0.93. These findings illustrate elevated precision, recall, accuracy, and F1-Score within the context of hybrid filtering, achieved through the application of Support Vector Machine for classification and the implementation of the SMOTEN technique.

Implementation of Continous Delivery using Jenkins And Kubernetes with Docker Local Images

Danur Wijayanto, Arizona Firdonsyah — Sun, 01 Oct 2023 00:00:00 +0700

The need for software has increased and the development process has also become more complex as technology has developed which resulted application will take a long time to deployed, which can be completed in weeks or even months. One reason is the large number of teams involved in application development, especially the Development Team and Operations Team. These teams can cause the application delivery process to the user's side to be hampered if a conflict occurs, so the term DevOps appears. Support for DevOps continues to be improved, so there is CI/CD (Continuous Integration and Continuous Delivery/Deployment). Banyak penelitian mengenai CI/CD yang sudah dilakukan dan menggunakan tools Jenkins, Kubernetes, dan Docker. Namun penelitian yang sudah ada menggunakan repository DockerHub untuk menyimpan Docker Image This research focuses on the local implementation of the docker image which is then run with Kubernetes Orchestration so it can reduce the platform used by CI/CD. This implementation requires conversion from docker image to Kubernetes Image Cache. The results show that the Continous Delivery implementation using Kubernetes with the Local Docker image is successful and can run well. The results show that the average time required to create a docker image is 649 seconds (10 minutes 49 seconds) and image conversion process which takes an average time of 275 seconds (4 minutes 35 seconds). Research can be developed further by researching techniques to shorten build time, minimize resource utilitation and minimize time conversion from docker images to Kubernetes Image Cache.

The Optimization of CNN Algorithm Using Transfer Learning for Marine Fauna Classification

Insidini Fawwaz, Yennimar, N P Dharsinni, Bayu Angga Wijaya — Sun, 01 Oct 2023 00:00:00 +0700

Marine fauna are all types of organisms that live in the marine environment. Marine fauna is also an important part of the marine ecosystem that has an important role in maintaining environmental balance. However, the survival of marine fauna is threatened due to activities carried out by humans, such as pollution, overfishing, industrial waste disposal into marine waters, plastic pollution and so on. Therefore, efforts are needed to monitor and protect marine fauna so that marine ecosystems can remain stable. One way to monitor marine fauna is by using classification technology. One of the technologies that can be used in marine fauna classification technology is Convolutional Neural Network (CNN). CNN is one of the classification methods that can be used to classify objects in images with a high level of accuracy. The CNN architecture models used are MobileNet, Xception, and VGG19. Furthermore, the method used to improve the performance of the CNN algorithm is the Transfer Learning method. The test results show that the MobileNet architecture model produces the highest accuracy value of 91.94% compared to Xception and VGG19 which only get an accuracy value of 87.64% and 88.42%. This shows that the MobileNet model has a more optimal performance in classifying marine fauna.

Android-based Automatic Steak Grilling Tool

Irma Salamah, Yunita Syaniah, Irawan Hadi — Sun, 01 Oct 2023 00:00:00 +0700

In an era of rapid technological development, technology is increasingly accessible and easily applied by humans. One of the significant developments is the Internet of Things (IoT), where physical devices such as sensors, equipment, and vehicles are equipped to communicate and interact via the Internet network. The application of IoT has expanded to various sectors, including culinary. In this regard, preparing and presenting food, especially steaks, becomes an exciting focus. There are multiple types of steaks, such as sirloin and tenderloin, and cooking involves various techniques, such as searing and grilling. However, suitability for maturity and risk during cooking is challenging for steak makers and connoisseurs. To overcome this, the application of IoT is needed in an automatic steak roaster to be a promising solution. This research is also equipped with real-time monitoring via an Android application. This aims to ensure proper doneness and consistent results in the steak cooking process. This research makes an automatic steak grill with a success rate of 83%, which shows that the tool's performance and functionality align with expectations. This tool also has an Android application to monitor and control the device remotely efficiently. This research gives confidence that this can be a solution that has been developed and provides significant benefits in roasting steaks with automatic monitoring and operation.

Comparison of LSTM and GRU Models for Forex Prediction

Mohammad Rezza Pahlevi, Kusrini, Tonny Hidayat — Sun, 01 Oct 2023 00:00:00 +0700

Trading foreign currencies worth trillions of dollars takes place daily in the forex market, characterized by highly volatile movements. The forex market operates on bid and ask prices, with exchange rates determined by the principles of supply and demand. Trading involves currency pairs like EUR/USD, where the value of the Euro is compared to the US Dollar, serving as a basis for analyzing price fluctuations. Due to the volatile nature of forex, market participants must make informed decisions when buying and selling, as improper choices can result in financial losses. One approach to mitigating risk in forex trading decisions is through the use of forecasting techniques. This research study employs LSTM and GRU methods to predict forex trends, which are evaluated using various dataset divisions. The most accurate results are obtained using a dataset of 4979, split into three equal parts: 80% for training, 10% for validation, and 10% for testing. This approach yields an RMSE value of 0.054, MAPE of 0.037, and R-square of 97%

The LSTM and Bidirectional GRU Comparison for Text Classification

Hannan Asrawi, Ema Utami, Ainul Yaqin — Sun, 01 Oct 2023 00:00:00 +0700

Although the phrases machine learning and AI are frequently used interchangeably and are frequently discussed together, they do not have the same meanings. While all artificial intelligence (AI) is machine learning, not all AI is machine learning, which is a key distinction. In the beginning, machine learning and natural language processing (NLP) are related since machine learning is frequently employed as a tool for NLP tasks. The advantage of NLP is that it can perform analysis, and examine a lot of data, including comments on social media accounts and hundreds of online customer evaluations. Text classification is essentially what needs to be done. This study compares Bidirectional GRU and LSTM as text classification algorithms using 20,000 newsgroup documents from 20 newsgroups from The UCI KDD Archive.

After using the suggested model, we compare it to the long short-term memory and bidirectional GRU models for accuracy and validation. The results of the two comparisons show that the bidirectional GRU model performs better than the long short-term memory model. And this is a successful classification of text using a deep learning algorithm that uses a bidirectional GRU.

Teacher Quality Affects On Graduation Of Study Programming Data Approach There With CRISP-DM Method

Sun, 01 Oct 2023 00:00:00 +0700

Each student's graduation is influential to the teacher in every subject that can be predicted based on the pattern of habits of the teacher who presents the subject. Web Proggramming is the subject of study that must be completed by every student. If this course is not completed, it is not allowed for the student to take other courses related to it. The custom patterns of teachers in this study were taken from 300 student respondents. An analysis is done to compare the results of questionnaire scores with the assessment of college admissions teachers. From the results of the comparison, it is possible to predict the graduation rate of students to the web programming course. The results of the experiment were that 72% of the students received highly influential predictions, 12% Influential, 7% Sufficient, 5% Influential and 4% Highly Influential.

Improving IT Support Efficiency Using AI-Driven Ticket Random Forest Classification Technique

Nathaniel Crosley, Ito Wasito — Sun, 01 Oct 2023 00:00:00 +0700

This research project aims to improve IT support efficiency at Indonesian company XYZ by using AI-based IT support ticket classification integration. This method involved collecting over 1,000 support tickets from the company's IT ticketing system, GLPI, and pre-processing the data to ensure the quality and relevance of the data for analysis. Claims data is enriched with relevant features, including textual information and categorical attributes such as urgency, impact, and requirement expertise. To improve the ticket preference matrix, AI-based language models, especially OpenAI's GPT-3, are used. These templates help to reclassify and improve the work of IT support teams. In addition, the ticket data is used to train the Random Forest classifier, allowing automatic classification of tickets based on their specific characteristics. The performance of the ticket classification system is evaluated using a variety of metrics, and the results are compared with alternative methods to assess effectiveness. of the Random Forest algorithm. This evaluation demonstrates the system's ability to correctly classify and prioritize incoming tickets. The successful implementation of this project at Company XYZ is a model for other organizations looking to optimize their IT support through AI-driven approaches. By providing simplified ticket classification and admission ticket reclassification based on AI algorithms, this research helps leverage AI technologies to improve IT support processes. Ultimately, the proposed solution benefits both support providers and users by improving efficiency, response times, and overall customer satisfaction.

Comparison of Sentiment Analysis Methods on Topic Haram of Music In Youtube

Rahmat Saudi Al Fathir As, Ema Utami, Anggit Dwi Hartono — Sun, 01 Oct 2023 00:00:00 +0700

Sentiment analysis on video lectures on YouTube that discuss the haram of music is an exciting topic to find out public opinion. This study aims to find what factors affect the model's accuracy in sentiment analysis, especially on video lecture content on YouTube. The data used is comment data on three video lectures that discuss the haram of music, which has been labelled into two categories: positive and negative. The data is divided into two categories, namely primary data, as many as 2099 data that have not been normalized, while secondary data has 1001 data that have been normalized. The experiment shows that the validity of the data, labelling the data, the amount of data, and preprocessing are essential points in forming a good sentiment analysis classification model because, from the test results, it was found that imbalance techniques such as SMOTE, word embedding word2Vec and FastText, and SVM and KNN classification algorithms do not provide maximum accuracy if the data used primary data. However, the data imbalance process, such as oversampling and SVM and KNN classification algorithms, will provide better model accuracy if used with secondary data. Based on the trial results, it is found that when using the SVM algorithm, primary data produces the highest accuracy at 58.35%, while secondary data is 72.23%. If using KNN, the primary data provides the highest model accuracy at 53.54%, while the secondary data has the highest accuracy at 72.81%. Based on these results, it was found that the validity of the data or data must be appropriate and related to the case raised and labelling the data must be done carefully because the most crucial is the inappropriate data in preprocessing the data must be done correctly, if data preprocessing is done in an inappropriate way then data imbalance techniques such as oversampling do not have enough influence on increasing accuracy, but if on the contrary then accuracy will increase. The selection of the right word embedding also affects accuracy. It is necessary to do many experiments to select the correct algorithm and follow the data owned because selecting the correct algorithm will provide maximum accuracy model results

Implementation of Transfer Learning in CNN for Classification of Nut Type

Sun, 01 Oct 2023 00:00:00 +0700

Nut has a high nutritional value and is widely used as an ingredient in cooking and snacks. Nut is included in the group of grains and has many types. Each type of nut has different nutritional content. Some types of nuts can also cause allergies or negative reactions in certain people, so it is important to identify the type of nut to be consumed. There are many types of nut that are different from each other, but some of them are similar. This makes it difficult to distinguish between the types of nuts, so there is a need for technology that can accurately identify nut types. Transfer Learning method is used to utilize trained models and applied to nut type classification. The two CNN models used are Inception V3 and Xception. The dataset consists of 11 types of nuts consisting of 1,320 data. The data is divided into 60% for training data and 40% for validation data. Preprocessing is done to ensure the image size is consistent and clarify the focus on the data image to be tested. The training results show that the Xception model is superior to Inception V3, with an accuracy of 86.36% on the validation data, while Inception V3 only reached 74.05%. Xception is able to predict nut types more precisely.

Satellite Images Classification using MobileNet V-2 Algorithm

Sun, 01 Oct 2023 00:00:00 +0700

Satellite imagery is an invaluable source of visual information for environmental monitoring and land mapping with high resolution and wide coverage. In this modern technological era, advances in Deep Learning technology have brought great benefits in utilizing satellite images for various purposes. One of the efficient Deep Learning models for satellite image classification is MobileNet V-2, which is specifically designed for devices with limited resources such as smartphones. This study aims to develop an accurate satellite image classification model using Convolutional Neural Network algorithm and MobileNet V-2 model. The data used is taken from the RSI-CB256 dataset developed through crowdsourcing data. This research resulted in the performance of three deep learning models, namely ResNet50, MobileNet V-2, and VGG-16. ResNet50 is the highest model performed best during the training phase, achieve an accuracy of 98.40%. MobileNet V-2 and VGG-16 followed with 95.64% and 96.62% accuracy, respectively. The evaluation results demonstrate the model's strong ability to accurately classify satellite imagery and strengthen the model's ability to generalize well. With high accuracy and the ability to run on smartphone devices, this model has the potential to provide valuable information for governments and scientists in preserving the earth and better responding to environmental changes.

Performance Comparison between Signature Cryptography: A Case Study on SNAP Indonesia

Moehammad Ramadhoni, Handri Santoso — Sun, 01 Oct 2023 00:00:00 +0700

SNAP (Standar Nasional OPEN API Pembayaran) was submitted by several sub-working groups formed jointly by ASPI and the Bank of Indonesia for encouraging digital transformation in the banking industry. In the document Pedoman Tata Kelola (Bank of Indonesia, n.d.), there is the use cryptographic algorithms that are used as validation for third parties to use the Open API. The algorithms used in the document are HMAC and RSA. The third party will send the signature in the API header along with the sent API payload. The signature describes the body payload, the endpoint URL that was called by the third party, and the time when the API call was made, so the signature will change all the time. However, there are other algorithms that can be used as a form of validation, such as ECC and ZK-SNARK. In this journal, the performance of the four cryptographic algorithms is compared. The performance we compare is overall speed when creating the signature and verifying it. The result is that HMAC is the most efficient algorithm, but for financial data, it is better to use ECC which uses asymmetric keys and is faster than RSA contained in the SNAP document, especially when 256 bits security level that ECC could be 10 times faster then RSA.

Performance of Various Naïve Bayes Using GridSearch Approach In Phishing Email Dataset

Rizki Rahman, Ferian Fauzi Abdulloh — Sun, 01 Oct 2023 00:00:00 +0700

The background is the increasing cybersecurity threats in the form of phishing attacks that can be detrimental to individuals and organizations. The purpose of this research is to compare the performance of four Naive Bayes variants in classifying phishing emails with a method that involves a data pre-processing stage, phishing emails are collected, cleaned, and converted into appropriate numerical features. Next, the GridSearch approach was used to find the best parameters. This research objective is to understand how each Naive Bayes variant works on phishing email datasets. This phishing detection task is based on the following performance evaluation criteria such as accuracy, precision, recall, and F1-score. In this study, Bernoulli got the best accuracy of 97.34% but when the results obtained a hyperparameter, the results showed an increase with the most optimal results and the best performance is Bernoulli 97.38%. The research results are to provide an in-depth insight into the effectiveness of each variant of Naive Bayes in dealing with phishing email datasets and researchers in selecting the most suitable Naive Bayes variant for phishing detection tasks. In addition, the applied GridSearch method can guide how to find the best parameters for Naive Bayes models in other contexts. In summary, this study focuses on analyzing the performance of four variants of Naive Bayes Gaussian, Multinomial, Complement, and Bernoulli with the best algorithms Bernoulli 97.38%.

Fingerprint Identification for Attendance Using Euclidean Distance and Manhattan Distance

Sun, 01 Oct 2023 00:00:00 +0700

Attendance is an action to confirm that someone is present at the office, school, or event. The use of attendance in an agency or company is really important as it can improve the level of discipline and productivity. However, the traditional way of doing attendance is considered less effective, less secure, and more difficult to organize. Therefore, a modern attendance system that utilizes fingerprints can be the right solution, especially because every fingerprint is unique. In this research, we focus on designing a fingerprint identification system for attendance purposes by using two distance measure methods, namely Euclidean Distance and Manhattan Distance. The dataset used in the research contains 111 fingerprint images with 90 images for training the designed fingerprint identification system and the remaining 21 images for testing the system. Each fingerprint image has undergone image pre-processing stage before being used. We compare Euclidean Distance and Manhattan Distance based on their performances in identifying fingerprint. From the test results, the fingerprint identification accuracy obtained using Euclidean Distance is 76.19%, while the accuracy obtained using Manhattan Distance is 71.43%. In general, both algorithms succeed in providing the correct identification results. This proves that Euclidean Distance and Manhattan Distance can be utilized for fingerprint identification purposes.

Prediction of Student Graduation with the Neural Network Method Based on Particle Swarm Optimization

Hafis Nurdin, Sartini, Sumarna, Yana Iqbal Maulana, Verry Riyanto — Sun, 01 Oct 2023 00:00:00 +0700

In private universities in Indonesia, student graduation is something that is worth paying attention to, because it will be an aspect of the success of the university. Universities certainly have data on students who graduated, where student graduation data is very important to be taken into consideration by private universities, however with a lot of data it will make it difficult for private universities to find information from this data. Other researchers have previously carried out student graduation data with the same data by examining student graduation data using other methods. So we need a data mining algorithm that has never been tested on student graduation data. The method used is the neural network method with an optimization method, namely the particle swarm optimization method, to test the data, which will later produce information that is very useful for universities. After testing the student graduation data and getting accuracy results using the neural network method of 84.55% and after being optimized using the particle swarm optimization method, the accuracy results were optimal with a value of 86.94%. These results can be used by private universities to predict that students will graduate on time before they take their final semester so that the student graduation rate will be high.

Ontology-Based Food Menu Recommender System for Patients with Coronary Heart Disease

Najla Nur Adila, Z. K. A. Baizal — Sun, 01 Oct 2023 00:00:00 +0700

Coronary heart disease is one of the leading causes of death. Knowledge of dietary patterns and proper food selection is an effort to address the risk and support coronary heart disease's healing process. Therefore, this study developed a food menu recommender system as a reference for patients with coronary heart disease. The recommender system is crucial in creating a proper dietary pattern for managing personalized meal plans. The system calculates the required nutritional needs of users. Ontology is used to represent knowledge about nutrition data and food intake. The ontology base with Semantic Web Rule Language (SWRL) enables the system to identify the most suitable foods for patients with coronary heart disease. We use SWRL rules to generate recommendation conclusions based on the existing ontology. Using this language enhances the descriptive logic capabilities, as the rules can overcome the limitations of the ontology language. Therefore, the system is built to find food menu options that match the required nutrition for patients. The nutritionist knowledge will be used to measure the system's performance compared to the recommendations made by nutritionists. From the user data sample, 150 recommended food menu data were obtained. The validation performance results obtained a precision 0.893, recall 1, and F_Score 94.3%.

Analysis of User Adoption Levels of JAKI Application Using the Government Adoption Model (GAM)

Zahra Anadya Kirani, Rio Guntur Utomo, Muhammad Faris Fathoni — Sun, 01 Oct 2023 00:00:00 +0700

This study delves into an analysis of the adoption patterns within the Jakarta Today e-government application (JAKI) through the dual lenses of the Government Adoption Model (GAM) and the Structural Equation Model (SEM). Encompassing JAKI users aged 17 years and above, the research encapsulates a substantial sample size of 384 individuals. The research findings underscore the pivotal role of key factors in driving e-Government adoption within the context of JAKI. Notably, Perceived Service Response, Perceived Trust, Perceived Uncertainty, Perceived Security, and Privacy collectively wield a significant and affirmative impact on the Adoption of e-Gov. However, intriguingly, factors including Perceived Awareness, Computer-self Efficacy, Availability of Resources, Perceived Ability to Use, Perceived Compatibility, Perceived Functional Benefit, Perceived Image, Perceived Information Quality, and Multilingual Option do not exert a notable influence on the Adoption of e-Gov. These insights proffer invaluable guidance for the Jakarta City Government, facilitating an enhanced understanding of user perceptions and needs. By meticulously addressing the determinative factors that engender a favorable adoption environment, the government stands poised to elevate the efficacy and reach of its e-government service, thus fostering greater citizen engagement and interaction with the JAKI application.

Image Augmentation for BreaKHis Medical Data using Convolutional Neural Networks

Maie Isthigosah, Andi Sunyoto, Tonny Hidayat — Sun, 01 Oct 2023 00:00:00 +0700

In applying Convolutional Neural Network (CNN) to computer vision tasks in the medical domain, it is necessary to have sufficient datasets to train models with high accuracy and good general ability in identifying important patterns in medical data. This overfitting is exacerbated by data imbalances, where some classes may have a smaller sample size than others, leading to biased predictive results. The purpose of this augmentation is to create variation in the training data, which in turn can help reduce overfitting and increase the ability of the model to generalize. Therefore, comparing augmentation techniques becomes essential to assess and understand the relative effectiveness of each method in addressing the challenges of overfitting and data imbalance in the medical domain. In the context of the research described, namely a comparative analysis of augmentation performance on CNN models using the ResNet101 architecture, a comparison of augmentation techniques such as Image Generator, SMOTE, and ADASYN provides insight into which technique is most suitable for improving model performance on limited medical data. By comparing these techniques' accuracy, recall, and overall performance results, research can identify the most effective and relevant techniques in addressing the challenges of complex medical datasets. This provides a valuable guide for developing better CNN models in the future and may encourage further research in developing more innovative augmentation methods suitable for the medical domain.

Decision Making for the Best Hospital Evaluation Using the AHP-COPRAS Method

Akmaludin, Adhi Dharma Suriyanto, Nandang Iriadi — Sun, 01 Oct 2023 00:00:00 +0700

The assessment of the Hospital is very important in providing a rating as a measure of accreditation. Hospitals should provide the best service to visiting patients. The purpose of this study is to provide a ranking of hospitals that should be a measure of the appropriateness of a hospital as seen from a number of criteria as a true assessment. The criteria used as a barometer for the assessment consist of ten criteria, namely Administrative Services, Doctor Services, Pharmaceutical Installation Services, Cleanliness, Convenience, Security, Number of Administrative Personnel, Number of Pharmaceutical Personnel, Administrative Complaints, and Pharmacy Installation Complaints. The scoring system that will be applied uses the Analytical Hierarchy Process (AHP) method as a determinant of the weight of each criterion and the Fuzzy assessment uses the Complex Proportional Assessment (COPRAS) method. The two methods will be collaborated as a determinant in the process of giving a rating system to the ten hospitals that are included in the priority assessment. The results obtained from the ranking process of the two AHP and COPRAS methods are seen from the acquisition of the utility value produced, the largest value obtained from the utility gives the best rating. The utility value is obtained from the total assessment of a number of criteria for each hospital and the largest utility rating does not exceed one. The highest value of the utility generated by the RS10 alternative with a utility scale of 0.098 as the best value.

Multidimensional Knapsack 0-1 Solution With Algorithm Evolution Pso-Ga

Yudistira Arya Sapoetra, Azwar Riza Habibi — Sun, 01 Oct 2023 00:00:00 +0700

This paper develops the particle swarm optimization (PSO) method and uses a genetic algorithm (GA) by changing the distribution of articles in the initialization of the initial position. PSO at this time the search and speed of particles will always go to the best solution so that by narrowing the search area will be faster by updating the best position of PSO. While the Genetic algorithm plays a role to get an expanded search area for PSO solutions by utilizing crossover and mutation in GA. So that GA will expand the range of candidates for the best solution in PSO. From each of the advantages of PSO Update and GA will be combined to get Evolutionary PSO-GA (EVPGA) that can minimize error and speed up computation (itation) in finding the best solution. By using the Multidimensional Knapsack data set, the results of EVPGA get an average speed of 24.9s with an error of 1.49%.

Comparison of NB and SVM in Sentiment Analysis of Cyberbullying using Feature Selection

Selamet Riadi, Ema Utami, Ainul Yaqin — Sun, 01 Oct 2023 00:00:00 +0700

In the past few decades, the internet has become an inseparable part of human life. It provides ease of access and permeates almost every aspect of human existence. One of the internet platforms that is widely used by people around the world is social media. Apart from being spoiled with the convenience and efficiency offered by social media to support daily life, it has gained popularity among a wide audience. This has positive implications when utilized effectively, but it cannot be denied that there are negative consequences if not utilized properly. One such consequence is the prevalence of cyberbullying activities on social media. Cyberbullying has become a major concern for the public and social media users, prompting researchers to leverage information technology in developing technologies that can identify the elements of cyberbullying, particularly on social media platforms. Sentiment analysis has been employed by researchers to identify the components of cyberbullying in online platforms. Sentiment analysis involves the application of natural language processing techniques and text analysis to identify and extract subjective information from text. This study aims to compare the Naive Bayes algorithm and the Support Vector Machine algorithm, while utilizing feature selection, specifically chi-square, to enhance the accuracy of both algorithms in classifying Instagram comments. The experimental results indicate that the Multinomial Naive Bayes (MNB) algorithm outperforms the Support Vector Machine (SVM) algorithm, achieving an accuracy of 83.85% without feature selection and 90.77% with feature selection. Meanwhile, SVM achieves an accuracy of 82.31% without feature selection and 90% with feature selection. Evaluation through the confusion matrix and classification report reveals that MNB exhibits better precision and recall rates compared to SVM in identifying bullying and non-bullying classes. The use of feature selection enhances the performance of both algorithms in classifying Instagram comments related to cyberbullying.

Ontology-Based Food Recommender System for Nutrition in School-Age

Dinda Atikah Wulandari, Z. K. A. Baizal — Sun, 01 Oct 2023 00:00:00 +0700

Nutrition plays an important role in the body and child development. Therefore, it is very important for parent to understand the nutritional needs of children to grow healthy and smart. If nutritional intake is not met, malnutrition can occur in children it interferes whit their growth and development process. The food recommendation system in this study is based on knowledge modeling. The focus of the research is to develop a recommendation system using ontology with Semantic Web Rule Language (SWRL) and form a knowledge base according to the guidelines proposed by Recommended Nutrient Intakes (RNI). Additionally, an Artificial Intelligence (AI) telegram chatbot named NutritionChildreBot was developed for this purpose. The recommended food menu is following the nutritional needs of children aged 7-9 years. The acquired knowledge base will be managed to provide information to users. The results of this research evaluation are in the form of recommendations for selecting foods that meet children’s nutritional needs based on information obtained from reliable sources.Based on this value, the calculation of precision, memory, and F_Score obtained is 97,9% of the accuracy of the results recommended by the system

A DESIGN UI/UX E-LEARNING ENGLISH MOBILE USING USER CENTERED DESIGN (UCD) METHOD

Dwi Rizky Alamsyah, Mochzen Gito Resmi, Irsan jaelani — Sun, 01 Oct 2023 00:00:00 +0700

E-learning English Mobile application design is a Mobile-based application design needed by SMAN 2 Purwakarta for future learning purposes. With this application design SMAN 2 Purwakarta has a picture of learning applications that will facilitate students and teachers in learning and teaching activities.

In this application design research, there are several methods that are widely used by previous researchers. One of the methods used in this research is the User Centered Design (UCD) method which is a new method in system development. UCD is a language that is widely applied in describing designs. The concept of UCD is the user as the center of the system development process, and the goals, the system environment are all based on the user experience.

The results of this study produced a learning application prototype, namely E-learning English Mobile. After testing using the System Usability Scale (SUS). The average value obtained is 78. It can be concluded that the design of the E-Learning English Mobile application is acceptable because it meets the Acceptable category

Densenet Architecture Implementation for Organic and Non-Organic Waste

Allwin M. Simarmata, Philander Salim, Netral Jaya Waruwu, Jessica — Sun, 01 Oct 2023 00:00:00 +0700

Garbage is the result left over from the process of daily human activities and activities which are considered no longer suitable for use, ranging from household waste to large-scale industrial waste. Therefore, the classification of waste is important because the problem of waste disposal is increasing and the way of processing is wrong. This research focuses on the classification of organic and non-organic waste using the DenseNet architecture. The dataset is processed first and each image in the dataset is resized to 128x128 pixels before being used in the model. We then trained all DenseNet types namely DenseNet121, DenseNet169, DenseNet 201, and compared their performance. Based on the test results, all DenseNet models that were trained were able to produce good accuracy, precision, recall, and F1 scores in garbage classification. In particular, our designed DenseNet121 model achieves 93.1 accuracy, 94.08% precision, 94.00% recall, 94.03% F1 score and 1min 34s training time as the best among other models. These results prove that the DenseNet architecture can be used to classify organic and non-organic waste correctly.

Optimization of Delay Using Killer Whale Algorithm (KWA) on NB-IoT

Sun, 01 Oct 2023 00:00:00 +0700

Abstract: NB-IoT is designed to connect IoT devices with low-power, wide-area coverage and efficient costs. Ensuring optimal data transmission delay is a challenge in NB-IoT implementation. Inadequate coverage can hinder IoT adoption. Optimization balances energy saving and delay trade-off. The Killer Whale Algorithm (KWA) optimizes delay by adjusting repetition variables. KWA addresses dimensions, variable limits. Applying KWA in NB-IoT optimizes transmission, enhancing QoS. Optimizing delay involves reducing latency in uplink data transmission using repetition variables. This study applies KWA to optimize NB-IoT delay. Analysis in Table 4 shows non-linear repetition-distance correlation. Interestingly, delay outcomes exhibit a contrasting relationship. Still, delay remains advantageous, remaining under 1 second even at 10 km, specifically 9.2674 ms (0.0092674 seconds). This thesis aims to optimize delay in NB-IoT network transmission using the Killer Whale Algorithm (KWA), crucial for modern communication networks and IoT applications. Leveraging KWA, the research identifies solutions to reduce transmission delay, enhancing efficiency and meeting IoT communication demands for speed and timeliness

Pneumonia Classification Based on Lung CT Scans Using Vgg-19

Sun, 01 Oct 2023 00:00:00 +0700

This research harnesses technology for critical health applications, specifically, pneumonia detection through medical imaging. X-ray photography allows radiologists to visualize the patient's health state, including the detection of lung infections signifying pneumonia. The study's centerpiece is the application of the VGG-19 model in classifying lung CT scan images, helping discern normal from pneumonia-indicative conditions. A comprehensive preprocessing procedure is employed, entailing pixel rescaling and data augmentation techniques. To address data imbalance, a critical issue in machine learning, we incorporate the Synthetic Minority Over-sampling Technique (SMOTE). The developed VGG-19 model demonstrates impressive performance, achieving a 94.6% accuracy rate in classifying lung CT scans. This finding underscores the potential of the VGG-19 model as a reliable tool for pneumonia detection based on lung CT scans. Such a tool could revolutionize the field, providing an efficient and accurate method for early pneumonia diagnosis, thereby allowing for timely treatment.

Virtual Space For Virtual Reality Exhibitions With Oculus Quest Devices

Muhammad Rusdi Rahman, M Suyanto, Dhani Ariatmanto — Sun, 01 Oct 2023 00:00:00 +0700

Based on information from the Ministry of Cooperatives and MSMEs, currently the number of MSMEs has reached 64.2 million euros and their share of GDP is 61.07% or 8,573.89 trillion rupiah (Coordinating Ministry for the Economy of the Republic of Indonesia, 2021). The contribution of MSMEs to the Indonesian economy includes the ability to absorb 97 percent of the current total employment and generate 60.4 percent of the total investment. (Ministry of Investment, 2021). VR (Virtual Reality) technology is a technology that allows users to feel in a virtual (virtual) world in visual form and users can interact with a virtual environment simulated by a computer in the form of Android. The focus of this research is a technical way of creating a virtual space or virtual space and 3D objects in displaying products from SMEs to be marketed to consumers with the virtual reality method using the oculus quest device, this research uses the Luther method, a six-stage process for creating multimedia that includes concept, design, material gathering, assembly, testing, and distribution. System testing was carried out using black box testing and usability testing using the SUS Score standardization, with a total of 47 respondents getting an average score of 54%. This average number exceeds 50% of the standard SUS Score Analysis, so the virtual reality exhibition space is categorized as suitable and OK for use by users. And it can also help MSMEs in carrying out virtual reality-based online marketing.

The Sentiment Analysis of BBCA Stock Price on Twitter Data Using LSTM and Genetic Algorithm Optimization

Rizki Tri Setiawan, Erwin Budi Setiawan — Sun, 01 Oct 2023 00:00:00 +0700

In today's business world, there is significant development and emergence of various and diverse innovations. Therefore, every company needs to develop itself in various ways, one of which is going public. This involves a company selling a percentage of its value to the public in order to facilitate its growth in every aspect required. However, it is not easy for issuers to attract investors to invest their capital because each investor has different criteria in terms of investment unit value. Essentially, the stock price depends on the strengths and weaknesses of the company. Hence, in order to expand the market and manage customer relationships, information is needed as a decision support. One of the sources of information that can be used is Twitter, which includes positive, neutral, and negative opinions. This study employs the LSTM classification method and word embedding using GloVe, followed by Genetic Algorithm optimization, which is used to predict sentiment in tweets related to the BBCA stock. The model is built with classification using Long Short-Term Memory to determine the level of accuracy produced. Then, the word embedding method using GloVe is used, and the obtained results with the GloVe-LSTM method yield an overall accuracy score of 71%. Furthermore, the optimization method using Genetic Algorithm is applied to enhance the previous method, GloVe-LSTM, resulting in an accuracy of 87% with the best individual values of 111,170, 0.398, 93, etc., and the best fitness score of 0.8724.

Innovative Role of Blockchain Pharmaceutical Supply Chain Digital Transformation: Enterprise Architecture Perspective

Bayu Yasa Wedha, Michael Sagar Vasandani, Alessandro Enriqco Putra Bayu Wedha — Sun, 01 Oct 2023 00:00:00 +0700

The advent of the Fourth Industrial Revolution has compelled numerous industries to undergo digital changes or transformations. One example is the pharmaceutical industry, which is responsible for providing medicinal products. The pharmaceutical supply chain assumes a crucial position within the pharmaceutical industry as it enables the secure, streamlined, and dependable transportation of medications from producers to individuals in need. The issue of prioritizing digital transformation within the pharmaceutical supply chain has emerged as a significant problem for hospitals and pharmaceutical businesses. Integrating different components inside the system is effectively supported by the substantial function fulfilled by Enterprise Architecture in this specific context. The objective is to minimize errors, enhance inventory management, optimize product distribution, and guarantee the safety and quality of pharmaceuticals. However, the process of adequately monitoring and verifying data has its challenges. However, these challenges can be efficiently addressed through implementing blockchain technology. In addition to this, Blockchain technology has the potential to enhance industrial efficiency. Utilizing blockchain technology enables the facilitation of transparency, immutability, and data integrity across the entirety of the supply chain. Integrating Enterprise Architecture electronic automation with Blockchain technology enables pharmaceutical enterprises to establish robust systems facilitated by Smart Contracts. The use of this system is expected to significantly enhance automation and regulatory adherence within supply chain processes, leading to notable advancements in operational efficiency, security, and data accuracy. Integrating blockchain technology and smart contracts enables pharmaceutical enterprises to enhance their product offers to hospitals and patients at reduced expenses, facilitating notable advancements in Enterprise Architecture.

Digital Transformation in University: Enterprise Architecture and Blockchain Technology

Iswahyudi, Djarot Hindarto, R. Eko Indrajit — Sun, 01 Oct 2023 00:00:00 +0700

Implementing digital transformation in higher education is now required for effective adaptation to rapid technological advances. Utilizing Enterprise Architecture (EA) with blockchain technology is a recommended strategic approach for implementing digital transformation. The college first ascertains the University's digital transformation requirements and goals, which include data management, security, operational efficiency, and transparency. In addition, formulating a strategic plan to determine the optimal integration of blockchain technology within the college's architectural framework, including the judicious selection of the most suitable blockchain platform, is essential. In addition, developing a company's architecture must prioritize seamless integration with existing systems while upholding data security and consistency principles. Consideration should be given to the significance of training and awareness building among faculty and students. In addition, the implementation process must be conducted in phases with consistent monitoring and evaluation. The success of this project is contingent upon the formation of partnerships and collaborations with blockchain technology companies, as well as a thorough understanding of the applicable regulatory framework. By adopting this methodology, the University can increase operational efficiency, bolster data security measures, and improve the educational experience. This research aims to increase efficiency, data security, transparency, and educational innovation in universities using Blockchain Technology and University Enterprise Architecture.

Expert System for Diagnosing Learning Disorders in Children Using the Dempster-Shafer Theory Approach

Murien Nugraheni, Rini Nuraini, Mursalim Tonggiroh, Siti Nurhayati — Sun, 01 Oct 2023 00:00:00 +0700

Learning disorders can occur in children where a child experiences difficulty mastering important skills such as reading, writing, or arithmetic. Learning disorders can have an emotional impact on children, such as low self-confidence, anxiety, or frustration. Therefore, it is important for parents and educators to recognize the signs of learning disorders so that appropriate intervention can be given. The aim of this research is to develop an expert system that can diagnose learning disorders in children using the Dempster-Shafer Theory algorithm to make it easier to diagnose and produce the right diagnosis. The Dempster-Shafer Theory approach has the ability to provide probability values in evidence based on the level of belief and reasoning in accordance with logic and then combine it with information from certain events. This research produces an expert system built on a website that can diagnose based on symptoms and display diagnosis results, definitions of types of learning disorders, and treatment options. The accuracy test results show a value of 92%, which means that the system built using the Dempster-Shafer Theory approach is able to diagnose learning disorders in children well.

Sentiment Classification of Fuel Price Rise in Economic Aspects Using Lexicon and SVM Method

Muhammad Fikri Alfauzan, Yuliant Sibaroni, Fitriyani — Sun, 01 Oct 2023 00:00:00 +0700

After being hit by COVID-19 for a long time around the world which resulted in the paralysis of all countries, especially the economic aspects of all countries that dropped dramatically, the world was again shocked by the conflict between Russia and Ukraine which resulted in an increase in world oil prices including in Indonesia, many people complained and opposed the government's policy of increasing fuel prices because fuel affects various aspects, including economic aspects. Based on these problems, researchers use sentiment analysis methods that aim to find out people's opinions on issues that are being discussed throughout Indonesia and this research focuses on comparing the SVM algorithm with TF-IDF feature extraction then using K-Fold Cross Validation after that it is compared with the Lexicon Inset dictionary, in this case the model with Lexicon Inset which contains weighting on each word. In this study, it was found that the dataset model using the SVM algorithm with TF-IDF feature extraction and then using K-Fold Cross Validation obtained an average accuracy of 0.85 using the SVM algorithm. While the model using the automatic labeling dataset using the Indonesian sentiment Lexicon (Lexicon Inset) obtained an average accuracy of 0.68. Classification using SVM with TF-IDF feature extraction is superior to using Lexicon Inset.

Comparison of RNN Architectures and Non-RNN Architectures in Sentiment Analysis

Djarot Hindarto — Sun, 01 Oct 2023 00:00:00 +0700

This study compares the sentiment analysis performance of multiple Recurrent Neural Network architectures and One-Dimensional Convolutional Neural Networks. THE METHODS EVALUATED ARE simple Recurrent Neural Network, Long Short-Term Memory, Gated Recurrent Unit, Bidirectional Recurrent Neural Network, and 1D ConvNets. A dataset comprising text reviews with positive or negative sentiment labels was evaluated. All evaluated models demonstrated an extremely high accuracy, ranging from 99.81% to 99.99%. Apart from that, the loss generated by these models is also low, ranging from 0.0043 to 0.0021. However, there are minor performance differences between the evaluated architectures. The Long Short-Term Memory and Gated Recurrent Unit models mainly perform marginally better than the Simple Recurrent Neural Network, albeit with slightly lower accuracy and loss. In the meantime, the Bidirectional Recurrent Neural Network model demonstrates competitive performance, as it can effectively manage text context from both directions. Additionally, One-Dimensional Convolutional Neural Networks provide satisfactory results, indicating that convolution-based approaches are also effective in sentiment analysis. The findings of this study provide practitioners with essential insights for selecting an appropriate architecture for sentiment analysis tasks. While all models yield excellent performance, the choice of architecture can impact computational efficiency and training time. Therefore, a comprehensive comprehension of the respective characteristics of Recurrent Neural Network architectures and One-Dimensional Convolutional Neural Networks is essential for making more informed decisions when constructing sentiment analysis models.

Blockchain-Based Academic Identity and Transcript Management in University Enterprise Architecture

Djarot Hindarto — Sun, 01 Oct 2023 00:00:00 +0700

The present research examines the implementation of Blockchain-based Identity and Academic Transcript Management in a university's enterprise architecture. This research is motivated by the increasing demand for secure, transparent, and efficient student identity management and the need to store easily verifiable academic transcripts. Blockchain technology has been spotlighted because it provides security and data integrity solutions. This research aims to determine if integrating Blockchain into the University's Enterprise Architecture can improve the management of student identities and academic transcripts by reducing the risk of forgery and facilitating more dependable access for interested parties. Increased security and efficiency in managing student data are the practical implications of this research, which can help universities reduce the risk of data loss and increase stakeholder trust. This research method includes surveying various universities that have adopted Blockchain technology in their academic identity and transcript management. In addition, we will assess its technical implementation, evaluate its effect on efficiency, and conduct interviews with university personnel involved in the implementation process. This study's anticipated outcome is providing universities planning to adopt Blockchain technology for Enterprise Architecture with actionable guidance. This research will identify the benefits, challenges, and best practices of integrating Blockchain in academic identity and transcript management and lay the groundwork for further improvement of educational services in university settings.

Embarking on Comprehensive Exploration of Classification System of Fruits and Vegetables

Bayu Yasa Wedha, Michael Sagar Vasandani, Alessandro Enriqco Putra Bayu Wedha — Sun, 01 Oct 2023 00:00:00 +0700

This research thoroughly investigates the fruit and vegetable classification system, emphasizing exhaustive investigation. This research aims to comprehend, analyze, and document numerous facets of this classification with MobileNetV2. This investigation included a comprehensive literature review, field investigation, and review of relevant scientific documents. In this investigation, we divide the classification of fruits and vegetables into various levels, ranging from the most general, such as kingdom and division, to the most specific, such as order and family. We also investigate the central role of taxonomy in comprehending the evolutionary and phylogenetic relationships between various categories of fruits and vegetables. This research enables us to identify and understand the taxonomic relationships between multiple varieties of fruits and vegetables and classify them into the appropriate botanical families. In addition, we investigate the global diversity of fruit and vegetable varieties, emphasizing the significance of conservation and genetic management to preserve the diversity of these precious commodities. In their efforts to comprehend, manage, and maintain the genetic variety of fruits and vegetables, this research provides researchers, botanists, and producers valuable insights. The findings of this study indicate that investigating fruit and vegetable classification systems is a crucial step in comprehending and conserving this irreplaceable natural resource, which provides direct benefits to humans in the context of global biodiversity conservation. MobileNetV2 research results accuracy epochs(5) = 94.84%, epochs(10) = 98.35%, epochs(15) = 98.69%.

Ambon Banana Maturity Classification Based On Convolutional Neural Network (CNN)

Sun, 01 Oct 2023 00:00:00 +0700

The banana (Musa paradical), is an excellent fruit produced nationally and high in vitamins. In Indonesia, banana production is at a higher level than other fruit products. However, one of them is the issue with bananas' post-harvest, which arises when they are produced in huge quantities on a large scale or by an industry that sorts bananas. So far, the determination of the maturity level of bananas is done by relying on visual analysis limited to the color of the skin by the human eye. However, this identification approach has several drawbacks. First, this method requires significant effort in the banana sorting process. In addition, the perception of the fruit's maturity level can vary, because humans can experience fatigue and lack of consistency in judgment. In addition, human judgment is also influenced by subjective factors that can affect the final result. Considering this problem, developed a system to classify the ripeness level of Ambon bananas. This system utilizes image enhancement features to increase contrast, which is implemented using a Convolutional Neural Network (CNN). The classification process is carried out through image processing using MATLAB R2022a software, which forms the basis of a classification system with 4 classes which include 486 images of unripe Ambon bananas, 235 images of half-ripe Ambon bananas, 309 images of perfectly ripe Ambon bananas, 184 images of rotten Ambon bananas. The dataset analyzed in this study totaled 1214 data divided into 1093 training data and 121 test data. The CNN method is used in this data classification, and the results show an accuracy rate of 95.87%.

Analysis of the Level of Public Satisfaction on the Tiktok Application as an E-Commerce

Deci Irmayani, Fransisco Alexander Sinaga, Masrizal — Sun, 01 Oct 2023 00:00:00 +0700

Abstract: Online shopping has become one of the alternatives used by society today. This happens because doing online shopping saves a lot of time. There are many online shops in Indonesia and are used by many people, such as Shopee, Lazada and Tokopedia. But now there is an application that initially only acted as a social media platform, but now also doubles as an E-commerce application, namely TikTok. TikTok has now become an E-commerce application. The prices given are also very cheap and there are lots of promotions given to customers. But there are still some people who don't want to shop online on TikTok on the grounds that the goods are not good. So from these 2 things, research needs to be made to determine the level of public satisfaction with the TikTok application as an E-commerce. The aim of this research is to see how many people are satisfied shopping on TikTok. This research was carried out using a classification model in data mining using the K-Nearest Neighbor (kNN) method and the Decision Tree method. The classification results obtained were 119 community data (for representation of 96.74%) and for people who were dissatisfied with the TikTok application as an E-commerce it was 4 community data (for representation of 3.25%). These results provide the conclusion that many people are satisfied shopping on the TikTok application as E-commerce

Detection And Classification of Citrus Diseases Based on A Combination of Features Using the Densenet-169 Model

M. Haikal Firdaus, Ema Utami, Dhani Ariatmanto — Sun, 01 Oct 2023 00:00:00 +0700

This research is motivated by the urgent need to improve the capability of detecting diseases in citrus plants, which play a crucial role in maintaining agricultural sector productivity. Diseases such as blackspot, canker, and greening can have a serious impact on harvest yields and overall plant health. Therefore, this research aims to enhance the accuracy in classifying diseases in citrus plants by applying a Deep Learning approach. In this study, we chose to adopt the DenseNet-169 architecture and conducted experiments with two different scenarios: one using original features and the other using a combination of features. This method was employed to classify four different classes, namely blackspot, canker, greening, and healthy plants, using an LDI dataset consisting of 3,000 images. This dataset was divided into three parts, namely training, testing, and validation sets. The experimental results indicate that the DenseNet-169 model with the use of feature combination achieved the highest accuracy rate at 96.66%, whereas the model using only original features achieved 91.33%. This significant improvement of 5.33% in accuracy provides strong evidence that the feature combination approach has a highly meaningful positive impact on the model's ability to identify and classify diseases in citrus plants. These findings confirm that the use of feature combinations is a highly effective strategy in improving the model's performance in disease classification tasks in citrus plants.

Three Pass Protocol for Key Security Using Affine Cipher Algortima and Exclusive-or (Xor) Combination

Muhammad Ikhwan Harahap, Suherman, Rahmat W Sembiring — Sun, 01 Oct 2023 00:00:00 +0700

Information is a very important concern in today's technological era, especially in terms of security through the exchange of information that is so fast that people can easily get various kinds of information. Information obtained easily through recording or openly disseminating data, XOR Cipher is an algorithm used to secure messages and texts but has weaknesses due to simple computation, therefore to strengthen security in XOR Cipher communication protocols can be added to secure key exchange on XOR Ciphers. Affine Cipher can be combined with Exclusive-OR (XOR) for text message security. Through the Affine Cipher algorithm with Key can change the unknown Plaintext, so that the Plaintext is kept secret. The Exclusive-OR (XOR) combination by changing each Character in each Plaintext according to the ASCII Code table can shorten the Key encoding process, so that Plaintext remains safe to send. The Affine Cipher Algorithm and the Exclusive-OR (XOR) Combination for Plaintext security levels are better because the Plaintext encryption and decryption processes are carried out twice with different cryptographic algorithms.

Identification of Public Library Visitor Profiles using K-means Algorithm based on The Cluster Validity Index

Sun, 01 Oct 2023 00:00:00 +0700

The existence of a public library in the Gampingan village has a positive impact, such as increasing the literacy culture of the village community. However, the library collection is not sufficient for the needs of visitors. Therefore, it is necessary to add library collections to fulfill the requirement. One of the solutions is mapping the library needs of visitors. The mapping can be done by identifying visitor profiles by grouping visitors based on the criteria of age, gender, type of visitor, and category of book library. One of the methods that can be used in the process of grouping visitors based on criteria is to use the K-Means Clustering method. Determining the number of K cluster centers at K-Means Clustering method that are not appropriate will give bad results, it is necessary to test the number of K cluster centers using the Cluster Validity index by measuring the clusters with cluster variance, within-cluster variance, and between-cluster variance. From the grouping process using K-Means Clustering with Cluster Validity index, we get 3 clusters of visitor profiles with a cluster variance value of less than 0.1. This shows that this method was able to identify the visitor profiles with high grouping accuracy values.

Optimizing Gender Classification Accuracy in Facial Images Using Data Augmentation and Inception V-3

Sun, 01 Oct 2023 00:00:00 +0700

In the digital era, facial recognition technology plays a crucial role in various applications, including gender classification. However, challenges such as variations in expressions and face positions, as well as differences in features between men and women, make this task formidable. This study aims to enhance the accuracy of gender classification using the Inception V-3 method and the Convolutional Neural Network (CNN), along with data augmentation techniques. The Inception V-3 method was chosen for its superiority in accuracy and speed. In contrast, the CNN model was selected in this study as a comparison and due to its algorithmic advantages in learning and extracting high-level features from images, including facial images, which are crucial for tasks such as gender classification. The data augmentation techniques in this study include rescaling, rotation, width and height shifts, shear range, zoom, horizontal flip, and fill method for model accuracy in gender classification with a small dataset. The study results indicate that the Inception V-3 model provides better accuracy (99.31%) in gender classification compared to the CNN model (81.31%). This conclusion underscores that the use of the Inception V-3 method with data augmentation techniques can improve the accuracy of gender classification in facial images.

Linear Regression Analysis To Measure The Correlation Between Poverty Rate And Stunting Rate

Suhaerudin, Ade Sumardi, Christina juliane — Sun, 01 Oct 2023 00:00:00 +0700

Children's stunting or growth disorders are becoming major global health issues, particularly in impoverished nations. It is characterized by short height for children and affects future economic potential, health, and cognitive development over the long run. Stunting has a detrimental effect on cognitive growth, schooling, and future economic production in addition to being a sign of dietary deficiencies. This study aims to analyze the relationship between poverty levels and stunting rates. Using secondary data from health surveys and population censuses, this study analyzed the rate of stunting in children aged 0-5 years and correlated it with poverty indicators at the household and community levels. Correlation analysis methods were used to assess the relationship between these variables, while controlling for confounding variables such as parental education, access to health services, and nutrition. The multiple linear regression test results prove that the incidence of stunting is influenced by the poor population variable by 34.1%, so there are other factors that influence it by 64.9%. The results of the analysis show that there is a significant positive correlation between the poverty rate and the prevalence of stunting. This finding underscores the importance of cooperation between the health and economic sectors in efforts to reduce stunting and poverty.

Comparison Of The C.45 And Naive Bayes Algorithms To Predict Diabetes

Alam, Divi Adiffia Freza Alana, Christina Juliane — Sun, 01 Oct 2023 00:00:00 +0700

Diabetes mellitus is an urgent global health problem and has a major impact on people around the world. This disease is characterized by high levels of sugar (glucose) in the blood due to disturbances in the production or use of the hormone insulin by the body. This study aims to carry out accurate early detection of diabetics so that they can be treated as soon as possible to reduce the risk of death and to compare the two algorithms that have the best level of accuracy. The algorithms used in this study are the C4.5 and Naïve Bayes Decision Tree Algorithms. The results of the experiments carried out in this study the Decision Tree Algorithm C4.5 and Naïve Bayes can be used in modeling the early detection of diabetes. The highest average accuracy results were obtained at 90.835% using the Decision Tree C4.5 Algorithm. As for the Naïve Bayes Algorithm, an average accuracy rate of 90.745% is obtained. The pruning process was carried out using the Decision Tree Algorithm C4.5, the accuracy performance increased to 91.30%. There were 18 patterns or rules for the early detection of diabetics from the built model. The determination of attributes, the number of attribute dimensions, and the number of samples greatly affect the performance of the model built.

PyTorch Deep Learning for Food Image Classification with Food Dataset

Iswahyudi, Djarot Hindarto, Handri Santoso — Sun, 01 Oct 2023 00:00:00 +0700

Classification of food images is crucial in today's increasingly digitally connected world. In the rapidly evolving mobile applications and social media era, the demand for an automated system that can recognize food types from an image is intensifying. This study employs deep learning and the PyTorch framework to develop a dependable and efficient solution for classifying food images. This research is motivated by the growing complexity of food introduction challenges. The primary challenge is improving the accuracy of food type recognition and overcoming variations in the visual presentation of food, such as lighting, shooting angles, and proportional and textural differences. Convolutional Neural Networks (CNN) are effective for image classification and are incorporated into the methods utilized. In addition, we employ ResNet101 transfer learning techniques to capitalize on the knowledge of trained models for large image datasets. The primary objective of this study is to develop a food image classification model that is accurate, training-efficient, and capable of accurately recognizing various types of food. In testing and evaluation, the developed model could realize multiple types of food with satisfactory accuracy. The accuracy of training reached 99.35%, while the accuracy of testing reached 94.65%. This study also reveals how Resnet101 transfer learning is utilized by deep learning technology.

Average Max Round Robin Algorithm: A Case Study

Tri Dharma Putra, Rakhmat Purnomo — Sun, 01 Oct 2023 00:00:00 +0700

Round Robin Algorithm is one wellknown algorithm in real time system. Several variants of round robin algorithms are in the fields. Average max round robin algorithm is a breakthrough to optimize context switching or also called quantum. Context switching is one big problem in round robin algorithm. To optimize high context switching is the key solution. This will make this algorithm efficient. There should be a way to optimize this context switching. Then the average max round robin algorithm is one solution to this problem. The average max algorithm is defined by finding the average of burst time then add the maximum burst time to the average burst time. Then calculate again the average of the two. The calculation will be iterated in the next round robin cycle. Here, in this journal, three case studies are discussed. Each with different burst times to understand this average max round robin algorithm more clearly. In the first case study we get turn around time 34 ms, and average waiting time 20.6 ms. In the second case study, we get average turn around time 21.8 ms, and the average waiting time 13 ms. And in the last case study, the third one, we get turn around time 12.2 ms, and the average waiting time 6.6 ms. There is no calculation for the second iteration for all case studies. Since the left burst time is only in one process. Optimizing the context switching, minimizing average turnaround time, and average waiting time is the key solution to round robin algorithm.

Enterprise Architecture Design for the Transformation of Online Financial Services

Bayu Yasa Wedha, Michael Sagar Vasandani, Alessandro Enriqco Putra Bayu Wedha — Sun, 01 Oct 2023 00:00:00 +0700

In the ever-expanding digital era, online financial services have emerged as a significant component of the global financial ecosystem. Digital transformation has altered the operations and consumer interactions of financial services companies. By implementing Enterprise Architecture design, financial services companies can adapt, flourish, and compete in an increasingly competitive environment. This article examines the essential function of Enterprise Architecture in the evolution of online financial services. We explore how Enterprise Architecture provides a strategic framework for driving change and innovation in financial organizations. We examine the primary elements of an Enterprise Architecture, such as the business model, technology infrastructure, and data. Then, we investigate how Enterprise Architecture improves operational efficiency and decreases risk in digital financial services. In addition, the significance of security and conformance in Enterprise Architecture for online financial services is examined. Enterprise Architecture is crucial in integrating security solutions and ensuring regulatory compliance in light of rising cybersecurity threats, strict regulatory compliance, and consumer demands for data privacy. We discuss how Enterprise Architecture facilitates improved consumer experiences in online financial services. By designing solutions with the consumer in mind, businesses can meet customer expectations, increase customer retention, and build a sustainable market share. Enterprise Architecture has become an indispensable instrument for successfully transforming online financial services. Through strategic planning, technology integration, data management, and a focus on security and customer experience, financial institutions can meet the challenges and seize the opportunities presented by the digital era.

User Interface Design for Baduy Ecotourism Website Using User Centered Design Method

Muhammad Farhan Fhalosa, Dawam Dwi Jatmiko Suwawi, Rosa Reska Riskiana — Sun, 01 Oct 2023 00:00:00 +0700

Baduy is one of the ecotourism destinations which offers captivating natural and cultural attractions, making it a worthy place to visit. The lack of available information in the media has piqued the author's interest in designing a visually appealing website to attract tourists. To address this issue, a user interface design will be created for the Baduy Tribe Ecotourism website using the User Centered Design (UCD) method. This method focuses on users and involves four stages, specifying the context of use, specifying requirements, designing a solution, and evaluating the design. Furthermore, to assess how well users interact with a product, usability testing will be conducted using the System Usability Scale (SUS) and Single Ease Question (SEQ). The usability testing results on the created website interface obtained a SUS score of 81 and generated 3 rating components in the SEQ method, fairly easy, easy, and very easy. Therefore, it can be concluded that the usability score falls within the category of good and is acceptable to users. Through this research, the author hopes that the user interface design for the Baduy Tribe Ecotourism website will meet users' needs, providing them with easy access to valid information and a seamless experience in discovering the natural wonders of the region.

Development of Learning Innovation Using Augmented Reality Technology with a Learning Management System as a Learning Supplement

Meyti Eka Apriyani, Budi Harijanto, Elok Nur Hamdana — Tue, 03 Oct 2023 00:00:00 +0700

Augmented Reality technology is making significant changes in the world of education by enriching the learning experience. This article discusses the development of innovative learning by integrating Augmented Reality into the Learning Management System as a learning support. This research aims to explain the concept of incorporating Augmented Reality into a Learning Management System as a complementary approach to increasing effectiveness and engagement in learning. The integration of Augmented Reality enables the presentation of learning content in a visual and interactive form, bridging the gap between theory and practice, and increasing student motivation and understanding. The results of using Augmented Reality in learning include increased understanding of the material, higher learning motivation, and the development of practical skills. However, this article also examines the challenges that may arise, both from a technical and pedagogical perspective, during the implementation of Augmented Reality in a Learning Management System. This research provides a comprehensive view of the potential of Augmented Reality as a complement to learning in facing the changing dynamics of modern learning. This article can be a guide for educational practitioners and researchers who are interested in incorporating Augmented Reality into learning contexts in various educational environments.

Sales Conversion Optimization Analysis Using the Random Forest Method

Kristiawan Nugroho, Th. Dwiati Wismarini , Hari Murti — Tue, 03 Oct 2023 00:00:00 +0700

Sales conversion is a challenging field of work in sales and business. Companies are competing to be winners by improving their services and hoping that their product sales can increase in various ways, including by using optimization theory. However, the lack of data analysis is a problem that is often encountered in optimizing sales conversions. Various machine learning-based methods have also been used to help analyze sales conversion optimization. This research uses the Random Forest method which is one of the more robust machine learning methods compared to other methods, namely Adaptive Booster (AdaBoost) and K-Nearest Neighbor (KNN) in analyzing sales conversion optimization. The results showed that the Random Forest method had the best performance in classifying data, by using the 10 cross validation technique the results were obtained with a Mean Squared Error (MSE) value of 0.928 and a Root Mean Square Error (RMSE) of 0.963, better than the Adaptive Booster method. and K-Nearest Neighbor which has lower performance. Sales conversion optimization processing using Random Forest is proven to have the best performance as evidenced by the small Mean Squared Error and Root Mean Square Error which means it has an accurate level of performance compared to other methods.

Information Systems UI/UX Design of Online Tickets for Situ Pasir Maung Tourism in Dago Village Using the Figma Application

Putri Eka Hidayanti, Rani Irma Handayani, Bakhtiar Rifai — Wed, 04 Oct 2023 00:00:00 +0700

has several interesting tours, one of which is Situ Pasir Maung, a place in the form of a natural tourism park located in Dago Village, Parung Panjang District, Bogor Regency. However, ticket purchases can only be made by buying directly on the spot when entering the tourist spot. This can make it difficult to order tickets due to the large number of visitors. So here a design for an e-ticket application will be made using the design thinking method to analyze and design a mobile application for online ticket ordering at Dago Tourism. In this design the editing software used is Figma, and in this study will only make UI/UX designs related to online ticket purchases. UI/UX design of the Design Thinking method for Situ Pasir. The Maung tourist ticket application was created and a prototype of the application was tested by sending a questionnaire to 20 respondents with an average score of 4.021 and most of the responses from potential users said that the tour ticket prototype was easy to understand and use. So here we will try to make an e-ticket application design. E-tickets can make it easier for buyers or visitors to get them because there is no need to come directly to tourist attractions. Ticket purchases can be made through easy-to-use online ordering

Bounding Box and Thresholding in Optical Character Recognition for Car License Plate Recognition

Wulida Rizki Sania, Christy Atika Sari, Eko Hari Rachmawanto, Mohamed Doheir — Thu, 05 Oct 2023 00:00:00 +0700

License plate recognition plays a central role in a variety of application contexts, including traffic management, automated parking, and law enforcement. Among the various approaches available, the Optical Character Recognition (OCR) technique has proven its effectiveness in recognizing characters in license plate images. This study describes an approach for detecting and recognizing vehicle license plates by utilizing the OCR method with Bounding Box, Thresholding, and template matching. In addition, this study uses MATLAB R2022a software as the main tool in developing and implementing the method. The goal is to recognize vehicle license plates from images, describe their characteristics, and generate relevant information. This approach involves a series of image processing steps starting with the pre-processing stage, followed by the process of binarization and license plate segmentation. After successfully isolating the license plate area, isolating the character using a bounding box is performed using image separation techniques. The OCR method is used to recognize license plate characters through comparison using the correlation method. Through a series of experiments on several image datasets, this approach succeeded in showing that out of 20 sampled license plate images, the results obtained were a reading accuracy of 93.55% of 100%, recognizing 13 out of 20 license plate images accurately when tested. Thus, the findings of this research are expected to contribute to the recognition of vehicle license plates that are accurate and efficient, by utilizing image processing techniques and OCR methods implemented using MATLAB R2022a software.

A Prototyping Model for Self-Appraisal Employee Performance Application Development in Cooperative

Shobrun Shobrun, Gia Anisa, Sinung Suakanto, Tien Fabrianti Kusumasari — Thu, 05 Oct 2023 00:00:00 +0700

Employee performance appraisal which is currently still being implemented cannot yet describe transparency. In addition, the process is still manual, which is far from effective. This study aims to develop a web-based employee performance appraisal application system that can increase the transparency of assessments in microfinance cooperatives. The focus is on using self-appraisal techniques to promote transparency in the appraisal process. Transparent assessments are critical to building trust and fairness during performance evaluations. Using the self-appraisal method, individuals can evaluate their own performance, skills, and achievements in a transparent and unbiased manner. This study investigates the process of making an employee performance appraisal system with the prototyping model method as a development method. The findings of this study contribute to existing knowledge about performance evaluation in microfinance services, particularly in relation to self-appraisal techniques, and offer practical insights for organizations wishing to increase the transparency of appraisals through web-based application systems.

Retracted: Geographic Information System for Customer Distribution Using the Haversine Algorithm

Bagus Setiawan, Samsudin — Sat, 14 Oct 2023 00:00:00 +0700

This paper was retracted, request by author

Information System for Monitoring Production Process of Dried Kelor Leaf Dried Using the FAST Method

Tue, 17 Oct 2023 00:00:00 +0700

Moringa or Kelor leaves, rich in nutrients and health benefits, are used in many culinary, supplement, and medicinal items. However, drying moringa leaves is a crucial step that impacts product quality. Companies must maintain product quality and production efficiency to meet rising demand. Since moringa leaf drying production management is difficult, this study uses the Framework for the Application System Thought (FAST) method. Its use in moringa drying allows thorough monitoring of temperature, humidity, drying duration, and other product quality factors. According to this research, using the FAST method in the moringa leaf drying production management monitoring application will help identify production issues, prevent product damage, and improve product quality. This research improves moringa production management and helps explain FAST method implementation in industrial process management. FAST is significant for monitoring applications because it can continually monitor all production conditions that affect drying moringa leaves. FAST can immediately detect dryer humidity issues. The FAST technique and moringa drying production management monitoring applications can be used to improve product quality, operational efficiency, and consumer safety in this research. Thus, this research gives tangible answers for the moringa processing business and can be applied to other industrial sectors facing comparable production process management issues.

Comparison of Activation Functions on Convolutional Neural Networks (CNN) to Identify Mung Bean Quality

Fri, 20 Oct 2023 00:00:00 +0700

Mung bean production levels by farmers in Indonesia are not stable. When there is a surplus, the stock of mung beans in the warehouse will accumulate, the storage factor affects the quality of mung beans. Indicators of quality mung beans can be seen from the color and size through direct observation. However, the aspect of view and assessment and the level of health of each observer is a human error in the classification of mung bean quality so that the results are less than optimal. One alternative way to identify object quality is to use deep learning algorithms. One of the popular deep learning algorithms is convolution neural network (CNN). This study aims to build a model to classify the feasibility of mung beans. The process of building the model also goes through the image preprocessing stage. In the process of building the model, there are ten setup parameters and four setup data used to produce the best model. As a result, the best CNN algorithm model performance is obtained from data setup I, with accuracy, precision, recall and F1 score above 75%. In addition, this study also analyzes Rel U and Adam activation functions on CNN algorithm on model performance in identifying mung bean quality. CNN algorithm with Adam activation function has 92% accuracy, 92.53% precision, 91.9% recall, and 92.19% F1 score. In addition, the performance of CNN algorithm with Adam activation function is superior compared to CNN algorithm with Adam activation function and previous study

Cloud Computing Analysis of Hybrid Networks on Raspberry

Risko Liza, Liza Fitriana, Junaidi, Daffa Maulana Siddiq — Sat, 21 Oct 2023 00:00:00 +0700

There are already a lot of cloud services on the internet which provide services, some of them even provide free facilities, but they are still limited in usage and capacity. Capacity is calculated based on saved files, temporary files, and even trash files. Moreover, data security cannot be guaranteed because the hardware is not properly set or owned by someone else. The cloud used cannot guarantee the connection and the bandwidth. The stigma of building cloud computing revolves around huge costs, not limited to operational and maintenance costs, nor the proper location for the cloud network equipment. As an effort to build data storage with large capacity, bandwidth regulation, and data protection, which is located in a private location, building a cloud computing service system which is efficient in time, cost, and place, and has good performance is no longer an impossible thing to do. With the help of Microcontroller technology, Raspberry Pi, a Cloud Computing with a Hybrid network could be built to reduce cost, time and space for the system. With the NextCloud application embedded into the cloud computing server, performance can be improved, including easy data synchronization that will flawlessly operate as a Client Server on a wide variety of today’s devices with examples being PCs, Tablets, Notebooks or Smartphones.

Exploring YOLOv8 Pretrain for Real-Time Detection of Indonesian Native Fish Species

Djarot Hindarto — Sat, 21 Oct 2023 00:00:00 +0700

The main objective of this research is to determine the efficacy of the YOLO model in detecting native fish species found in Indonesia. Indonesia has a variety of maritime natural resources and shows significant diversity. This research utilizes the YOLO architecture, previously trained on several datasets, for fish detection in the environment in Indonesian waters. This dataset consists of various fish species native to Indonesia and was used to retrain the YOLO Pretrain model. The model was evaluated using test data that accurately represents Indonesian water conditions. Empirical findings show that the modified YOLO Pretrain model can accurately recognize these fish in real-time. After utilizing YOLO and Pre-Train with Ultralytics YOLO Version 8.0.196, the results show an accuracy of 92.3% for head detection, 86.9% for tail detection, and an overall detection accuracy of 89.6%. The fish image dataset, consisting of a total of 401 images, is categorized into three subsets: the training dataset, which consists of 255 images; the validation dataset, which includes 66 images; and the testing dataset, which contains 80 images. This research has great potential for application in fisheries monitoring, marine biology research, and marine environmental monitoring. A real-time fish detection system for the Identification and tracking of fish species is carried out by researchers and field workers. The findings of this research provide a valuable contribution to ongoing efforts aimed at conserving marine biodiversity and implementing more sustainable management practices in Indonesia.

Hair Disease Classification Using Convolutional Neural Network (CNN) Algorithm with VGG-16 Architecture

Ichwanul Muslim Karo Karo, Dedy Kiswanto, Suvriadi Panggabean, Adidtya Perdana — Wed, 25 Oct 2023 00:00:00 +0700

Hair diseases are common and can be caused by a variety of factors, including genetics, stress, nutritional deficiencies, as well as exposure to sunlight and air pollution. Accurate diagnosis of hair diseases is important for proper treatment, but can be challenging due to overlapping symptoms. The development of the healthcare world has widely utilized machine learning and deep learning approaches to assist in the healthcare field. This research aims to develop hair disease classification using Convolutional neural network (CNN). The CNN-based approach is expected to help health professionals diagnose hair diseases accurately and provide targeted treatment. This research involves an experimental design with three main stages: identifying the research problem, conducting a literature review, and collecting data. The research uses a dataset of hair disease images obtained from Kaggle, which are annotated and organized based on different hair disease types. After the image data is collected, the image dataset will go through the image preprocessing stage. Experiments were conducted using hair disease image data with 15 epochs on a CNN Deep Learning model with VGG-16 architecture, and resulted in an accuracy of 94.5% and a loss rate of 18.47%, with a testing epoch time of 9 hours 48 minutes. The results of this study show that CNN with VGG-16 architecture can successfully classify 10 types of hair diseases

Mobile Apps-Based Cosmetic Equipment Selection Decision Support System Use Simple Additive Weighting (SAW) Method

Clarina Monica Destin, Yuli Asriningtias — Fri, 27 Oct 2023 00:00:00 +0700

Currently, more and more types of cosmetic products are appearing on the market, thus making cosmetic users confused in choosing cosmetics that suit their skin type and usage. Not a few people who use cosmetics wrongly which has a lot of bad effects on their face and has to be touched up repeatedly. This application facilitates users with 5 types of skin types, lip types and types of use that can be selected to make it easier for users to choose cosmetics that suit their skin type and usage. The application was built with a Decision Support System (DSS) using the Simple Additive Weighting (SAW) method to make calculations for selecting the most appropriate Make Over cushion and lipstick product for the user. In this application, the user can enter their skin type and then the system will calculate using the SAW method to get the Make Over cushion and lipstick that is most suitable for the user. The final result of the system will display the Make Over cushion and lipstick and their description with the highest calculated value to the user. Based on testing, this application is able to display cushion and lipstick products whose compatibility with users reaches 98%.

Enhancing Road Safety with Convolutional Neural Network Traffic Sign Classification

Djarot Hindarto — Thu, 02 Nov 2023 00:00:00 +0700

Recent computer vision and deep learning breakthroughs have improved road safety by automatically classifying traffic signs. This research uses CNNs to classify traffic signs to improve road safety. Autonomous vehicles and intelligent driver assistance systems require accurate traffic sign detection and classification. Using deep learning, we created a CNN model that can recognize and classify road traffic signs. This research uses a massive dataset of labeled traffic sign photos for training and validation. These CNN algorithms evaluate images and produce real-time predictions to assist drivers and driverless cars in understanding traffic signs. Advanced driver assistance systems, navigation systems, and driverless vehicles can use this technology to give drivers more precise information, improving their decision-making and road safety. Researcher optimized CNN model design, training, and evaluation metrics during development. The model was rigorously tested and validated for robustness and classification accuracy. The research also solves real-world driving obstacles like illumination, weather, and traffic signal obstructions. This research shows deep learning-based traffic sign classification can dramatically improve road safety. This technology can prevent accidents and enhance traffic management by accurately recognizing and interpreting traffic signs. It is also a potential step toward a safer, more efficient transportation system with several automotive and intelligent transportation applications. Road safety is a global issue, and CNN-based traffic sign classification can reduce accidents and improve driving. On filter 3, Convolutional Neural Network training accuracy reached 98.9%, while validation accuracy reached 88.23%.

Battle Models: Inception ResNet vs. Extreme Inception for Marine Fish Object Detection

Djarot Hindarto — Thu, 02 Nov 2023 00:00:00 +0700

Within the domain of deep learning applied to computer vision, there exists a significant emphasis on the competition between two prominent models, namely Inception ResNet and Xception, particularly in the field of marine fish object detection. The present study conducted a comparative analysis of two advanced neural network architectures in order to assess their efficacy in the identification and localization of marine fish species in underwater images. The two models underwent a rigorous evaluation, utilizing their capabilities in feature extraction. The findings indicate a complex performance landscape, wherein Inception ResNet exhibits remarkable accuracy in identifying marine fish objects, while Xception demonstrates superior computational efficiency. The present study elucidates the inherent trade-off between precision and computational expenditure, offering valuable perspectives on the pragmatic ramifications of choosing one model over another. Furthermore, this research underscores the significance of carefully choosing a suitable model that aligns with the particular requirements of object detection applications in the context of marine fish. This study endeavors to guide professionals and scholars in marine biology and computer vision, enabling them to make well-informed choices when utilizing deep learning techniques to detect maritime fish objects in underwater settings. The research specifically focuses on the comparison between Inception ResNet and Xception models.

Performance Comparison ConvDeconvNet Algorithm Vs. UNET for Fish Object Detection

Djarot Hindarto — Thu, 02 Nov 2023 00:00:00 +0700

The precise identification and localization of fish entities within visual data is essential in diverse domains, such as marine biology and fisheries management, within computer vision. This study provides a thorough performance evaluation of two prominent deep learning algorithms, ConvDeconvNet and UNET, in the context of fish object detection. Both models are assessed using a dataset comprising a wide range of fish species, considering various factors, including accuracy of detection, speed of processing, and complexity of the model. The findings demonstrate that ConvDeconvNet exhibits superior performance in terms of detection accuracy, attaining a noteworthy degree of precision and recall in identifying fish entities. In contrast, the UNET model displays a notable advantage in terms of processing speed owing to its distinctive architectural design, rendering it a viable option for applications requiring real-time performance. The discourse surrounding the trade-off between accuracy and speed is examined, offering valuable perspectives for algorithm selection following specific criteria. Furthermore, this study highlights the significance of incorporating a diverse range of datasets for training and testing purposes when utilizing these models, as it significantly influences their overall performance. This study makes a valuable contribution to the continuous endeavors to improve the detection of fish objects in underwater images. It provides a thorough evaluation and comparison of ConvDeconvNet and UNET, thereby assisting researchers and practitioners in making well-informed decisions regarding selecting these models for their specific applications.