Sinkron : jurnal dan penelitian teknik informatika

Design of Intelligent Model for Text-Based Fake News Detection Using K-Nearest Neighbor Method

2024-12-10T06:22:00+00:00

Text-based fake news detection is a crucial issue considering its negative impacts on society and individuals. One of the main impacts that has a significant and detrimental impact on society is disinformation, where false or misleading information can cause confusion and uncertainty in society. This can lead to misunderstandings and develop into riots in society which can lead to legal problems that are detrimental to society. In order to overcome this problem, a method is needed to detect fake news. This study aims to build a fake news detection method using machine learning, which is a technology widely used by researchers to detect and analyze past data. Various methods have been produced using machine learning, including the K-Nearest Neighbor (K-NN) method which is proposed as an effective solution to detect fake news. K-NN is a machine learning algorithm that works by classifying text based on its proximity to known data in feature space. This method is proposed because of its ability to handle non-linear data and its low complexity. The application of K-NN can increase the accuracy in detecting fake news by utilizing the characteristics of relevant text, thus helping in efforts to filter information and maintain the integrity of news circulating in the community. In a study conducted using the FakeNewsDetection dataset, the model evaluation results showed that KNN produced a Mean Absolute Error (MAE) of 0.011 and a Root Mean Squared Error (RMSE) of 0.077, better than the performance of other methods such as SVM and Neural Network.

Predicting IT Incident Duration using Machine Learning: A Case Study in IT Service Management

2024-12-18T05:10:17+00:00

In the digital era, ensuring customer satisfaction with IT services is crucial for business success. However, the complexity of IT infrastructure makes it difficult to manage services, requiring companies to focus on improving efficiency and reducing operational costs. One of the strategies used is Information Technology Service Management (ITSM), the main component of which is incident management, which aims to minimize service disruptions. While various studies on ITSM exist, research focused on Machine Learning models for predicting incident resolution times is relatively limited. This research aims to develop an incident resolution duration prediction model using a Random Forest Regressor-based regression approach. The dataset used is an event log from the ServiceNow system containing data on 24,918 incidents. The model was evaluated using the Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R2 metrics, where the model achieved a MAE of 14.33 hours, RMSE of 69.8 hours, and R2 of 0.98. These results show that the model can provide accurate predictions and support better decision-making in IT incident handling. Time-related features, such as sys_update_month and closed_month, proved to be the most influential factors in predicting incident resolution duration.

Detection of Plastic Bottle Waste Using YOLO Version 5 Algorithm

2024-11-18T05:27:43+00:00

Plastic bottle waste management has become one of the most pressing environmental issues, especially in countries with high plastic usage rates, such as Indonesia. This research uses the YOLOv5 (You Only Look Once version 5) algorithm to detect plastic bottle waste automatically. The YOLOv5 algorithm was chosen because it has efficient detection performance and high accuracy in small object recognition. The dataset consists of 500 images of plastic bottles obtained through cameras and internet sources. The data is processed through several stages: annotation (bounding box and labeling using Roboflow), split dataset (70% for training, 20% for testing, and 10% for validation), pre-processing (resizing images to 460x460 pixels), and augmentation (adding data variations to improve model performance). Training and evaluation of the YOLOv5 model using the precision metric of 89.8% indicates the ability of the model to accurately identify plastic bottles from the overall prediction, recall of 83.1% indicates the success of the model in detecting the majority of plastic bottles in the test data, and mean average precision (mAP) of 89.2% represents the average precision at various prediction thresholds. Test results on varied bottle image test data obtained detection accuracy between 82%-93%, indicating the model can recognize plastic bottles consistently. Sometimes, this model needs help detecting overlapping picture objects. However, this research proves the potential of the yolov5 algorithm as an automated litter detection solution that will be integrated with a system and support faster and better plastic waste management.

Analyzing PEGASUS Model Performance with ROUGE on Indonesian News Summarization

2024-12-10T06:20:46+00:00

Text summarization technology has been rapidly advancing, playing a vital role in improving information accessibility and reducing reading time within Natural Language Processing (NLP) research. There are two primary approaches to text summarization: extractive and abstractive. Extractive methods focus on selecting key sentences or phrases directly from the source text, while abstractive summarization generates new sentences that capture the essence of the content. Abstractive summarization, although more flexible, poses greater challenges in maintaining coherence and contextual relevance due to its complexity. This study aims to enhance automated abstractive summarization for Indonesian-language online news articles by employing the PEGASUS (Pre-training with Extracted Gap-sentences Sequences for Abstractive Summarization) model, which leverages an encoder-decoder architecture optimized for summarization tasks. The dataset utilized consists of 193,883 articles from Liputan6, a prominent Indonesian news platform. The model was fine-tuned and evaluated using the Recall-Oriented Understudy for Gisting Evaluation (ROUGE) metric, focusing on F-1 scores for ROUGE-1, ROUGE-2, and ROUGE-L. The results demonstrated the model's ability to generate coherent and informative summaries, achieving ROUGE-1, ROUGE-2, and ROUGE-L scores of 0.439, 0.183, and 0.406, respectively. These findings underscore the potential of the PEGASUS model in addressing the challenges of abstractive summarization for low-resource languages like Indonesian language, offering a significant contribution to summarization quality for online news content.

Enhanced Semarang Batik Classification using MobileNetV2 and Data Augmentation

2024-12-10T06:24:04+00:00

Batik, an Indonesian cultural heritage recognized by UNESCO, faces challenges in pattern identification and documentation, particularly for the younger generation. Previous studies on batik classification have shown limitations in handling small datasets and maintaining accuracy with limited computational resources. This research proposes an enhanced classification approach for Semarang Batik motifs using MobileNetV2 architecture combined with strategic data augmentation techniques. The study utilizes a dataset of 3,020 images comprising 10 distinct Semarang Batik motifs, implementing horizontal flipping, rotation, and zoom transformations to address dataset limitations. Our methodology incorporates transfer learning through ImageNet pre-trained weights and custom layer modifications to optimize the MobileNetV2 architecture for batik-specific features. The model achieves 100% accuracy on validation data, with precision, recall, and F1-scores consistently above 0.98 across all classes. The confusion matrix analysis reveals minimal misclassification between similar motif patterns, particularly in the Batik Blekok Warak and Batik Kembang Sepatu classes. This research contributes to cultural heritage preservation by providing an efficient, resource-conscious solution for automated batik pattern recognition, potentially supporting educational and commercial applications in the batik industry.

Leveraging Label Preprocessing for Effective End-to-End Indonesian Automatic Speech Recognition

2024-11-28T13:25:22+00:00

This research explores the potential of improving low-resource Automatic Speech Recognition (ASR) performance by leveraging label preprocessing techniques in conjunction with the wav2vec2-large Self-Supervised Learning (SSL) model. ASR technology plays a critical role in enhancing educational accessibility for children with disabilities in Indonesia, yet its development faces challenges due to limited labeled datasets. SSL models like wav2vec 2.0 have shown promise by learning rich speech representations from raw audio with minimal labeled data. Still, their dependence on large datasets and significant computational resources limits their application in low-resource settings. This study introduces a label preprocessing technique to address these limitations, comparing three scenarios: training without preprocessing, with the proposed preprocessing method, and with an alternative method. Using only 16 hours of labeled data, the proposed preprocessing approach achieves a Word Error Rate (WER) of 15.83%, significantly outperforming the baseline scenario (33.45% WER) and the alternative preprocessing method (19.62% WER). Further training using the proposed preprocessing technique with increased epochs reduces the WER to 14.00%. These results highlight the effectiveness of label preprocessing in reducing data dependency while enhancing model performance. The findings demonstrate the feasibility of developing robust ASR models for low-resource languages, offering a scalable solution for advancing ASR technology and improving educational accessibility, particularly for underrepresented languages.

Clustering Analysis of Stunting Risk Factors Using K-Means and Principal Component Analysis: A Case Study in Indonesian Regency

2024-12-13T07:45:42+00:00

Stunting, characterized by impaired growth and development in children, is one of the most serious public health problems often caused by chronic malnutrition. This study aims to identify patterns among stunting cases through clustering analysis of child health data. The algorithm used in this research uses K-Means. The dataset used in this study uses health data from 599 children in the Sambas Regency area of East Kalimantan Province. This dataset has several features that are quite diverse such as height, weight, age, nutritional intake, socioeconomic status, and others. This research process begins with cleaning the data, as well as looking at the correlation between features. One of the methods used is to conduct a data analysis process using Principal Component Analysis (PCA) which aims to reduce the dimensions of the data. After that, the process of finding the number of clusters using the Elbow method is carried out to determine the optimal number of clusters. This research uses 4 clusters in the process. The clustering results revealed that family structure (main family vs extended family) and parental income levels significantly influence stunting prevalence in the region.

Maturity Level Analysis of SPBE Service Domain Using Capability Maturity Model Integration at the Kominfo Palembang City

2024-12-10T06:23:27+00:00

This journal aims to analyze the maturity level of the implementation of Electronic-Based Government Systems (EBS) in the service domain at the Communication and Information Technology Office of Palembang City. This research uses the Capability Maturity Model Integration (CMMI) approach to evaluate processes, identify weaknesses, and provide recommendations for improvement. CMMI was chosen because it can measure process effectiveness and help organizations achieve optimal performance. The research was conducted using survey and interview methods to collect data related to SPBE implementation. The collected data was analyzed using the CMMI framework to determine the maturity level from level 1 (Initial) to level 5 (Optimizing). The results of the analysis show that the maturity level of the SPBE service domain at the Communication and Information Technology Office of Palembang City is at level 3 (Defined) with a maturity level value of 3.66 from a recapitulation of a value mapped to each process area: OPF, OPD, MA, CAR and PPQA. Some areas need to be improved, especially related to lack of clearly defined and consistently applied standard operating procedures (SOPs) leads to variations in service delivery and hampers the overall effectiveness of SPBE implementation, performance monitoring mechanisms such as tracking and evaluation of service delivery outcomes are inadequate which makes it difficult to assess the effectiveness of SPBE services, existing systems are not fully compatible or lack the necessary features to support technology integration within the SPBE framework leading to inefficiencies and failure to leverage technology to improve public services. This research contributes by providing strategic recommendations to improve the maturity of SPBE implementation at the Communication and Informatics Office of Palembang City. The recommendations given are increasing the capacity of human resources, consistent application of standard operating procedures (SOPs), and the use of more integrated technology to support more effective and efficient services. The results of this study are expected to serve as a guide for the Communication and Information Technology Office of Palembang City in implementing SPBE more optimally.

Median-Average Round Robin (MARR) Algorithm for Optimal CPU Task Scheduling

2024-08-27T07:26:39+00:00

Abstract: In operating systems, multitasking or multiprocessing terms are used. If more than one task operating consecutively, but the users feel that they are running simultaneously, than it is called multitasking. Round robin algorithm is a noted algorithm in multitasking. Several modifications of classical round robin algorithm have been proposed by experts. The idea behind these modifications are to get lower turnaround time and lower waiting time. The main topic’s discussion is about median-average round robin (MARR) algorithm. In this algorithm, the processes are arranged in ascending order. Then we get the median of the burst time. Afterwards, calculation of the average burst time is done. The summation of average and median, divide by two is the time quantum. So, the time quantum will be dynamic, based on each iteration of round robin. First iteration can have different time quantum compared to the second and so on. Each iteration will have one time quantum. Three analysis’s are given. Each with five processes. In the first analysis, time quantum for 1^st iteration is 11 and the 2^nd iteration is 4. The average turnaround time is 29. The average waiting time is 19. For the second analysis, time quantum for 1^st iteration is 10 and the 2^nd iteration is 8. The average turnaround time is 24.2. The average waiting time is 13.6. For the third analysis, time quantum for 1^st iteration is 10 and the 2^nd iteration is 9. The average turnaround time is 23.2. The average waiting time is 12.8.

Comparison of ARIMA and GRU Methods in Predicting Cryptocurrency Price Movements

2024-11-30T16:40:25+00:00

This study compares the effectiveness of the ARIMA and GRU models in predicting Bitcoin price movements, addressing the need for reliable predictive tools amidst the high volatility of the cryptocurrency market. Previous research has highlighted the strengths of each model in financial forecasting: ARIMA for short-term, stationary data and GRU for capturing complex temporal patterns. The purpose of this study is to evaluate which model performs better in the context of Bitcoin price prediction, offering insights for investors to minimize risks and enhance decision-making in this unpredictable market. The research methodology involves applying both models to Bitcoin price data and comparing their accuracy using the Mean Absolute Percentage Error (MAPE) across various forecasting intervals. Results indicate that GRU achieves higher accuracy in long-term forecasts, while ARIMA performs optimally for shorter time frames. However, both models demonstrate limitations, especially as the prediction horizon extends, underscoring the inherent challenges of cryptocurrency price forecasting. These findings suggest that GRU may be better suited for longer investment horizons, while ARIMA remains effective for short-term predictions. The conclusions affirm the potential of using these models selectively to align with specific investment strategies in cryptocurrency markets, although further research is recommended to improve predictive accuracy under evolving market conditions.

Stock Price Prediction Using TCN-GAN Hybrid Model

2024-11-23T01:15:31+00:00

The stock market plays a vital role in national economies, offering significant profit opportunities for investors while exposing them to substantial risks due to market uncertainties. Stock prices often experience significant fluctuations, making accurate prediction a challenging task. Temporal Convolutional Network (TCN) and Generative Adversarial Network (GAN) are the deep learning method proposed for this research. The purpose of this research is to analyze how well the TCN-GAN model predicts stock prices. Previous researches show both TCN and GAN perform well on time series data. TCN excels in analyzing time-series data while GAN enhances training by generating realistic simulations. By combining the strength of both models, this approach aims to enhance stock price prediction accuracy. The proposed model uses TCN as the generator within the GAN framework and a Multilayer Perceptron (MLP) as the discriminator. TCN handles the prediction task and is trained using the GAN model. The model is trained over 500 epochs, with a learning rate of 0.0004 for the generator and 0.0001 for the discriminator. During each epoch, the generator is updated twice to enhance its performance. The resulting model achieves a MAPE score of 2.16% and an RMSE score of 814.25 on the testing dataset, demonstrating excellent performance in stock price prediction despite significant price variations.

Parameter Testing on Random Forest Algorithm for Stunting Prediction

2024-11-28T13:26:33+00:00

Stunting is a significant public health problem, especially in developing countries like Indonesia. It is often caused by chronic malnutrition in the first 1,000 days of life, which can impact a child's physical growth and cognitive development. To find risk factors and find effective solutions, data analysis was conducted by utilising machine learning to predict stunting. This research uses the Random Forest algorithm with a focus on setting parameters such as n_estimators, max_depth, and the number of features to optimise model efficiency and accuracy. Using the 2023 Indonesian Health Survey data consisting of 25,800 data, this study managed to get the highest accuracy of 91.65% by a combination of Random Forest with parameter settings n_estimators 200, max_depth 30, and Synthetic Minority Oversampling Technique (SMOTE). Despite the high accuracy results, there are limitations such as potential noise coming from synthetic data from SMOTE and the limited number of features analysed. It is hoped that this research can contribute to the field of machine learning model development that is practically used to predict stunting, and support the government's efforts to reduce the stunting prevalence rate to 14% as targeted. This model also provides strategic insights for policy makers to design more effective data-driven interventions, which can help in decision making.

Comparison of RNN and LSTM Algorithms Based on Fasttext Embeddings in Sentiment Analysis on the Merdeka Mengajar Platform

2024-12-07T01:26:45+00:00

As of 2024, the Merdeka Mengajar Platform has been used by more than 3.5 million teachers across Indonesia. This number represents an increase of more than 3.85% compared to the previous academic year, which was 3.37 million. However, the utilization of this application has not yet reached the expected target number of users, so an analysis is needed to identify the factors causing this. This research uses Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) to perform sentiment analysis on reviews of the Merdeka Mengajar platform. RNN and LSTM are chosen for their advantages in handling sequential data, particularly in text processing for sentiment analysis. This research aims to address the challenges in understanding the positive or negative sentiments of users on the platform. The research methodology includes important stages such as data cleaning, preprocessing, and transforming text into numerical vectors using FastText embedding. Next, RNN and LSTM models are applied to predict sentiment based on patterns in the text data. The research results show that the LSTM model is capable of capturing long-term relationships in sequential data with an expected accuracy of 93.58%. Meanwhile, the RNN model yields a lower accuracy of 91.70%. The LSTM model is more effective in classifying sentiment with high accuracy, especially in text data with complex temporal contexts. This research contributes to understanding user perceptions and feedback regarding the Merdeka Mengajar platform, which is expected to provide insights for platform developers to enhance service quality.

Implementation of LSA for Topic Modeling on Tweets with the Keyword ‘Kemenkeu’

2024-12-10T06:26:53+00:00

This research explores public discourse on financial policies by analyzing tweets mentioning the keyword 'Kemenkeu' (Ministry of Finance). Using Latent Semantic Analysis (LSA), the study examined 10,099 tweets to uncover key topics that reflect public sentiment toward the Ministry’s policies. Preprocessing steps, such as stopword removal and stemming with Sastrawi, were essential to ensure the effectiveness of the analysis. The results revealed three main topics: Finance and Budget, Salaries and Employee Welfare, and Excise and Customs Regulations. These insights provide a better understanding of public opinion on financial issues and highlight the importance of proper text preprocessing in topic modeling. This approach demonstrates how LSA can be used as a tool for analyzing large-scale social media data, offering valuable input for policymakers. Future research could expand on this by using more advanced models or larger datasets to gain deeper insights.

Usability Evaluation of Lecturer Information System in ITB STIKOM Bali

2024-12-12T06:55:21+00:00

This research aims to provide a deeper understanding of the usability aspects of information systems, as well as help create more effective and efficient solutions in supporting academic activities in higher education. The ITB STIKOM Bali, Lecturer Information System (Sistem Informasi Dosen/ SID) is a system developed to assist lecturers in carrying out their academic responsibilities. This system has become a very vital tool in supporting various academic activities of lecturers. This research will be conducted using the Concurrent Think Aloud (CTA) method, Performance Measurement, and System Usability Scale (SUS) to assess effectiveness, efficiency, and user satisfaction of the system. The system was found effective, with a success rate exceeding 78%. Advanced users achieved a 95% success rate, while beginner users achieved 86%, with errors primarily in navigation-related tasks. User satisfaction analysis via SUS showed skilled users rated the system at 84.75 (Grade A, Acceptable), whilst beginner respondents scored 52.5 (Grades D, Marginal Low), reflecting usability challenges for beginners. Performance Measurement highlighted issues with small font sizes and unclear navigation, while CTA identified difficulties with the logout button, lack of search functionality, unreadable interface text, and unclear functional position menus. Recommendations include increasing font size to Arial 14, redesigning the logout button, adding search bars, and enhancing functional menus to include research and community service options. These improvements aim to enhance system usability and user experience across all proficiency levels.

Thyroid Disease Prediction Using Random Forest with KNNImputer for Missing Values

2024-12-14T19:04:08+00:00

Thyroid disease is a health dysfunction that requires immediate and accurate diagnosis. This research seeks to design a classification model based on the Random Forest algorithm to detect the type of thyroid disease utilizing data from the UCI Repository. In the data processing stage, KNNImputer is used to handle missing data by calculating the average value of the nearest neighbors based on Euclidean distance, thus ensuring better data quality for model training. The developed model was evaluated utilizing the confusion matrix, which showed an accuracy of 98%, with precision, recall, and F1 score values reached 98% based on weighted avg.These results corroborate that the proposed model is highly reliable in detecting various types of thyroid diseases, such as Negative, Hypothyroid, and Hyperthyroid. This research makes an important contribution to the application of data mining technology for medical diagnosis, while proving that optimal data processing through methods such as KNN Imputer can significantly improve model performance.

Optimizing Twitter Sentiment Analysis on Tapera Policy Using SVM and PSO

2024-11-12T03:45:29+00:00

This study aims to analyse the sentiment of Twitter users towards the Public Housing Savings (Tapera) policy in Indonesia using the Support Vector Machine (SVM) algorithm optimised by Particle Swarm Optimization (PSO). In recent years, social media has emerged as a primary platform for individuals to express their views and opinions on public policies. The government programme, Tapera, which was designed to increase access to housing for the public, attracted considerable attention, with a range of responses, including both positive and negative sentiments. The methodology employed in this study comprised the collection of data from Twitter, the processing of text, and the application of SVM-based classification techniques, reinforced by PSO, with the objective of enhancing the accuracy and efficiency of the model. The results demonstrated that the PSO-optimised SVM model exhibited an accuracy of 85%, accompanied by an Area Under Curve (AUC) value of 0.84 and a ROC curve that indicated the model's notable capacity for differentiating between positive and negative sentiments. These findings indicate the existence of certain sentiment patterns that can be utilised for the evaluation and improvement of Tapera policies. In conclusion, this research is expected to provide a comprehensive picture of the public response to the Tapera policy and present an analytical model that can be applied to evaluate other policies. Further research is recommended to expand data coverage and develop algorithms to achieve more accurate results.

Implementation of Case-Based E-Consultation to Handle Student’s Stress Levels

2024-11-16T09:09:17+00:00

Despite the declaration of the COVID-19 pandemic's end, the effects of some cases persist in the new normal era of 2023. Several cases indicate a decline in the learning motivation of students and university students, which significantly affects aspects of understanding, creativity, productivity, and learning outcomes. University students transition from learners who, during their high school years, spent more time studying online without directly interacting with peers or teachers. One of the causes of university student dropouts is internal issues due to students' inability to adapt to the university environment. The purpose of this research is to compile cases frequently experienced by university students that cause stress and lead to the decision to discontinue their studies. This is done to find solutions and prevent similar incidents from recurring. The implementation of e-counselling helps provide solutions in the form of action recommendations on how to address student issues. We conducted the research in several stages, including data collection, literature review, modelling, model evaluation, and prototype building and testing. We obtained the solution to the collected cases from the counsellor through a focus group discussion (FGD). This research employs case-based reasoning, utilizing four reasoning processes: retrieve, reuse, revise, and retain. We chose the modified weighted average similarity function to measure the case's similarity value with the cases in the case base. Through the case-based e-counselling system, the calculation results reveal the similarity between the new case and the old cases, recommending actions that counsellors have validated as valid solutions.

Embedded Smart Farming System for Soil and Hydroponic Planting Media Based on The Internet of Things

2024-11-25T05:33:17+00:00

Smart agricultural technology by applying the internet of things (IoT) purposes to make farmers' work more efficient due to the automation system and assist farmers in monitoring the condition of their agricultural land. The focus of discussion in this research is the application of smart agriculture system technology that uses the concept of embedded systems for soil and hydroponic planting media. This system applies an automation system for water irrigation and fertilizer irrigation using four tanks, namely a water source, a water irrigation tank, a fertilizer tank, and a water circulation system in hydroponics. The system is also equipped with weather monitoring based on temperature, rainfall, and light intensity. Other parameters contained in this system are soil pH, water pH, TDS, fertilizer availability, and irrigation pump status. The monitoring system based on the Android application displays all parameters and the status of the devices used.

Usability Evaluation of GetContact Application Using Post-Study System Usability Questionnaire and Retrospective Think Aloud

2024-11-28T13:35:06+00:00

GetContact, aplikasi manajemen dan proteksi panggilan spam dengan lebih dari 700 juta unduhan di Google Play Store, masih memiliki ruang untuk peningkatan kegunaan. Studi ini bertujuan untuk memanfaatkan kegunaan GetContact berdasarkan ulasan Quora, komentar Google Play Store, dan wawancara dengan pengguna di area Palembang. Metode yang digunakan adalah Post-Study System Usability Questionnaire (PSSUQ) dan Retrospective Think Aloud (RTA). Hasil PSSUQ dari 190 responden menunjukkan tingkat kegunaan keseluruhan yang baik dengan skor 2,73. Skala Kegunaan Sistem mencatat nilai 2,60, dan Kualitas Informasi mencapai 2,80, menunjukkan kegunaan yang memuaskan. Namun, kualitas antarmuka dengan skor 2,89 masih perlu ditingkatkan. Temuan dari metode RTA juga mengidentifikasi kendala dalam fitur dan antarmuka aplikasi. Studi ini menyimpulkan bahwa meskipun tingkat kegunaan GetContact secara keseluruhan baik dan diterima oleh pengguna, pengembangan lebih dari antarmuka dan fitur masih diperlukan untuk meningkatkan kegunaan secara keseluruhan dan menciptakan pengalaman pengguna yang lebih baik.

Sentiment Analysis of Tokopedia App Reviews using Machine Learning and Word Embeddings

2024-12-03T17:42:40+00:00

Tokopedia, a prominent e-commerce platform in Indonesia, generates vast amounts of user feedback, offering valuable insights into customer satisfaction through sentiment analysis. However, sentiment analysis of app reviews specifically on Tokopedia reviews remains underexplored. Sentiment analysis, also known as opinion mining, categorizes user sentiments into positive or negative, offering insights into user preferences and service shortcomings. While traditional text representation techniques like TF-IDF are widely used for sentiment analysis, they lack the semantic richness provided by advanced word embeddings such as Word2Vec and FastText, which excel at capturing contextual relationships between words. These methods have shown potential to enhance the performance of machine learning models in sentiment analysis tasks. This study evaluates the performance of three machine learning algorithms—Support Vector Machine (SVM), Random Forest (RF), and Gaussian Naïve Bayes (NB)—combined with Word2Vec and FastText feature extraction. A dataset of 59,811 Tokopedia app reviews was scraped from the Google Play Store, preprocessed, and subjected to feature extraction using Word2Vec and FastText. SVM achieved the best performance, with an accuracy of 89.06% using FastText and 89.02% using Word2Vec. RF ranked second with accuracies of 88.07% for FastText and 88.14% for Word2Vec. NB showed the lowest performance, achieving 84.26% with Word2Vec and 83.73% with FastText. Differences in performance between Word2Vec and FastText embeddings were minimal across all algorithms, highlighting their comparable effectiveness. These results underscore SVM’s consistent superiority across various configurations for sentiment analysis.

Transforming Real Estate: Leveraging TOGAF ADM for Digital Optimization in Enterprise Architecture

2024-11-08T03:31:36+00:00

In this research paper, we propose an Enterprise Architecture (EA) design for PT XYZ, a middle up class real estate development company in Indonesia, leveraging the TOGAF ADM framework. The study centers on optimizing five key business processes—commercial leasing, residential sales, hotel banquet rentals, waterpark ticket sales, and parking fee collection—to enhance operational efficiency and support digital transformation. Using ArchiMate modeling for clear visualization, this architecture spans from the Preliminary Phase, Phase A Architecture Vision, Phase B Business Layer, Phase C Information System Architecture (Application Layer) to the Phase D Technology Architecture. It provides a strategic blueprint to address common challenges like data fragmentation, reliance on manual processes and human resources readiness. By implementing this EA, PT XYZ can expect improvements in scalability, flexibility, and overall agility. This approach aims to position PT XYZ as a modern, digitally-driven entity, aligning technology investments with business objectives for long-term success. Future research is recommended to explore later phases of TOGAF ADM (Phase E – Phase H) and potentially integrate additional business areas for a holistic digital transformation.

Development of Mobile Application by Applying Content-Based Filtering

2024-12-13T07:41:08+00:00

The rapid advancements in information technology have transformed modern lifestyles, driving changes in consumer behavior and expectations, especially in the retail industry. This study focuses on developing a mobile application for Ampu Mart, a newly established retail business in Indonesia, to optimize product recommendation systems using the Content-Based Filtering (CBF) approach. The research integrates CBF with string matching and cosine similarity algorithms to provide personalized product recommendations based on customer preferences, enhancing user satisfaction and supporting more efficient purchasing decisions. The methodology involves several stages, including problem identification through observation and interviews, data collection on product attributes and customer preferences, system design, prototype development, implementation, and testing. The application leverages advanced algorithms to analyze product characteristics, ensuring relevant recommendations by matching user preferences with product attributes. User Acceptance Testing (UAT) conducted with 30 participants—customers, administrators, and management—evaluated the application's functionality, usability, accuracy, and performance. Results indicate that the mobile application improves the shopping experience and boosts sales by offering accurate, user-centered recommendations. The findings highlight the strategic importance of integrating intelligent technology into e-commerce platforms to enhance competitiveness in the retail market. Future work recommends incorporating Collaborative Filtering techniques to further enrich the recommendation system by analyzing collective customer behavior.

KNN Approach to Evaluating the Feasibility of Using Scientific Publications as Final Projects

2024-12-23T20:24:58+00:00

This study aims to explore the feasibility of using scientific publications as a substitute for traditional final assignments in higher education by applying the K-Nearest Neighbors (K-NN) algorithm. Traditional final assessments, such as theses, are widely used in evaluating students, but with the increasing availability of peer-reviewed scientific publications, there is potential to use them as a more dynamic and relevant assessment tool. This study uses a dataset containing scientific publications and theses, with features such as research quality, relevance, methodology, and clarity. This study applies the K-NN algorithm to classify these materials and determine whether scientific publications can serve as an effective substitute. The results show that the K-NN algorithm, using k=4, achieved 95% accuracy, successfully distinguishing between scientific publications and theses. However, some misclassifications occurred, indicating areas for improvement, such as incorporating additional features such as citation counts or peer-review scores. These findings suggest that scientific publications, if properly classified, can indeed replace traditional final assignments, encouraging critical thinking and engagement with current research. Future research should refine the feature set and explore other machine learning models to improve accuracy. The practical implications of this research are the potential to develop more innovative and relevant approaches to assessment in higher education, which are more aligned with modern educational practice.

Fraud Detection in Mobile Phone Recharge Transactions Using K-Means and T-SNE Visualization

2024-12-13T07:51:54+00:00

The surge in digital transactions has introduced vulnerabilities in mobile recharge systems, making them susceptible to fraudulent activities that compromise financial security and operational integrity. This study presents to address these challenges by employing a novel fraud detection framework that integrates K-Means clustering and t-Distributed Stochastic Neighbour Embedding (t-SNE) visualization. This work advances the field by integrating scalable, unsupervised learning techniques with robust visualization tools, offering a practical framework for fraud detection in mobile recharge systems. Leveraging a dataset of over 200,000 transactions, this research systematically identifies anomalies indicative of fraudulent behaviour, focusing on key transactional attributes such as processing times, geographic patterns, and error frequencies. The methodology begins with data preprocessing to ensure consistency, followed by the application of K-Means clustering to partition transactions into meaningful clusters. To enhance interpretability, t-SNE visualization is employed, enabling a clear representation of high-dimensional data and the identification of anomalous patterns. A comparative analysis with Autoencoders highlights the strengths of K-Means in terms of computational efficiency, interpretability, and clustering quality, as evidenced by higher Silhouette Scores (0.6215) and lower Davies-Bouldin Index values (0.7074). The combination of K-Means and t-SNE enables service providers to identify fraudulent activities with greater precision, offering actionable insights to mitigate financial risks. This study not only addresses the critical need for robust fraud detection systems but also lays a strong foundation for future advancements through the integration of hybrid models and enhanced feature engineering, demonstrating its adaptability to similar domains.

Evaluation of Clustering Algorithms for Identifying Shoe Characteristics Patterns at XYZ Footwear

2024-12-13T18:41:50+00:00

As the third-largest shoe-exporting country in the world, Indonesia faced a 25% decline in shoe exports in 2023 compared to the year before, both in terms of net weight and sales value. This decline in shoe exports occurred due to the increase of complexity and variety in customer orders to shoe manufacturers. These reasons require shoe manufacturers to enhance their production planning systems to become more efficient and competitive. To address this problem, this study explores the application of clustering algorithms to optimize the production planning process in shoe manufacturing companies. Using a case study at XYZ Footwear, clustering algorithms such as K-Means, Support Vector Clustering (SVC), and Deep Autoencoder were evaluated and compared to find the most effective algorithms in identifying patterns in shoe characteristics, thereby improving shoe manufacturers' production planning process. The datasets consist of the 2024 production season data, categorized into shoe categories, models, and variants, and purchase orders. The result shows that the combination of Deep Autoencoder and K-Means has better performance than just K-Means or Support Vector Clustering (SVC), achieving a silhouette score of 0.4822 and a Davies-Bouldin Index (DBI) of 0.6741. These findings highlight the effectiveness of combining deep learning (Deep Autoencoder) with clustering algorithms (K-Means) in identifying patterns in shoe characteristics.

Performance Level Analysis On Learning Vector Quantization And Cohonen Algorithms

2024-12-13T06:16:05+00:00

Biometric identification is an alternative for a security system that consists of physiological characteristics and behavioral characteristics. Physiological characteristics are relatively stable physical characteristics such as fingerprints, hand lines, facial features, tooth patterns, and the retina of the eye. Behavioral characteristics such as signature, speech patterns, or typing rhythm. The function of a signature is proof in a document which states that the party signing, knows and agrees to all the contents of a document. There are several stages in the signature pattern image recognition system, namely the signature pattern image is produced through a scanning process, then the resulting digital signature image is cut (scaling) manually, the next process is thresholding, edge detection, image division, and representation. input value. The method used in recognizing signature patterns is the learning vector quantization (LVQ) artificial neural network method and kohonen self-organizing map (SOM). In Learning vector quantization, the initial weights are updated using existing patterns. Meanwhile, in the self-organizing map method, Kohonen takes initial weights randomly, then these weights are updated until they can classify themselves into the desired number of classes. The processes that occur in the artificial neural network method require a relatively long time. This is influenced by the large number of data samples used as a means of updating the trained weights. From the results of the research conducted, it shows that the learning rate value that was built around 0.2 < α ≤ (10) ^ (-2) can produce better signature pattern recognition accuracy.

Analysis of Social Assistance Donor Classification at the Muhammadiyah Medan Orphanage Using SVM

2024-12-07T01:29:05+00:00

The Putra Muhammadiyah Orphanage in Medan City is a social institution that relies on donor assistance to support various social programs. The problem that occurs at the Putra Muhammadiyah Orphanage in Medan is the difficulty in identifying potential and non-potential donors who have the potential to provide sustainable social assistance contributions. This study aims to conduct a comprehensive analysis and classification of donors using the Support Vector Machine method, an effective method in machine learning to handle classification problems with SVM with high accuracy. The research data consists of donor data with several main characteristics such as the amount of donation, the frequency of donations given, and the type of assistance. The data is processed through a preprocessing stage including data normalization and data division into training and testing data. Then, the SVM model is trained to classify donors into two categories, namely Potential Donors and Non-potential Donors. Based on the data obtained from the donation bookkeeping records of the Putra Muhammadiyah Orphanage in Medan City, it can be concluded that around 55 potential donors out of 90 donors and 35 non-potential donors out of 90 donor data. From the results of the analysis and testing of the model conducted, it can be concluded that the SVM method can classify "Potential Donors" and "Non-Potential Donors" with a fairly high level of accuracy. The level of accuracy obtained reached up to 89% with a precision value of 93%, a recall value of 89% and an f1-score of 90%. With these results, this study can provide significant benefits in the management of social assistance, especially helping orphanages to determine who are potential and non-potential donors. Therefore, this study is expected to have an impact on improving the sustainability of social programs at the Putra Muhammadiyah Orphanage in Medan City.

Prediction of Organic Waste Deposits in Compost Houses using LSTM and ARIMA Algorithms

2024-12-02T10:27:30+00:00

Indonesia faces a significant waste problem and is becoming a global challenge, mainly due to inadequate food waste management. In Kendal District, the Environmental Agency struggles to optimize waste collection and predict the volume of organic waste. To address this issue, this study explores the application of predictive technology and data analysis to improve the efficiency of waste management. Two predictive models, ARIMA and Long Short-Term Memory (LSTM), were developed and compared by collecting historical data from Kendal Organic Compost House from 2020-2024 while for train and test data using data from January 2, 2023, to December 30, 2023. The ARIMA model showed better accuracy, capturing stable trends and seasonal patterns in the time series data, with an MSE of 72,799.49. Meanwhile, the LSTM model, although capable of handling non-linear and complex patterns, performed poorly with an MSE of 54,711,498,631,770.58, indicating a failure to accommodate sharp fluctuations in the data. These findings highlight the suitability of ARIMA for data with low volatility and strong seasonality, making it more reliable for short-term predictions. The results of this study are expected to assist the Kendal District Environmental Agency in planning efficient waste management strategies, optimizing compost house operations, and improving resource allocation. Future research should focus on the integration of external variables, such as weather and population dynamics, and explore hybrid models for better prediction.

Food and Physical Activity Tracking Application with Simple Dietary Pattern Analysis

2024-11-30T12:49:27+00:00

This study focuses on the development of a mobile application to track food intake and physical activity while offering simple dietary pattern analysis. The primary goal was to create an intuitive tool enabling users to log meals, record physical activities, and receive actionable feedback on caloric balance. Developed using Agile methodology, the application includes user-friendly interfaces for data entry, a dashboard for visualizing caloric intake and expenditure, and feedback to enhance users’ understanding of dietary habits. Results from a one week user testing phase demonstrated high user satisfaction, with participants appreciating the app’s simplicity and clarity in presenting health-related insights. The app effectively encouraged users to engage with their dietary and activity habits, promoting informed lifestyle decisions. However, limitations such as the lack of detailed macronutrient tracking and integration with wearable devices were identified, which could improve accuracy and broaden the app's appeal. Future improvements are suggested, including the addition of macronutrient analysis, wearable device compatibility, and features like goal-setting and gamification to enhance engagement. These findings indicate that a straightforward, user-friendly health tracking app can significantly increase health awareness and support behavior change, particularly for individuals new to health monitoring. The research highlights the potential of simple digital tools to foster sustainable health improvements while addressing users’ needs effectively.

Digital Transformation of Electricity Bill Collection: Predicting Delays Using Machine Learning

2024-12-21T11:15:00+00:00

Delays in electricity bill payments pose a significant challenge for PLN in maintaining financial stability and delivering equitable service quality to the public. This study aims to develop a payment delay prediction system to assist PLN UP3 Makassar Utara in prioritizing invoice distribution to customers with a high likelihood of late payments. The Random Forest algorithm was chosen for its ability to handle complex data and produce reliable predictions. This research analyses historical electricity customer data from 2018 to 2023, encompassing 227,163 entries. The data is processed using validation techniques such as K-Fold Validation and Rolling Window Validation to ensure the accuracy and generalizability of the model. The study's findings demonstrate that an accurate payment delay prediction model can be developed using the Random Forest method, incorporating historical features such as lag features, moving averages, and seasonal variables. Additionally, the system prioritizes invoice delivery to high-risk customers based on risk scores derived from historical delay patterns. This system reduces payment arrears at PLN UP3 Makassar Utara through proactive measures such as early notifications, personalized reminders, or payment incentives to encourage timely payments. As a result, the study indicates that the system effectively enhances the efficiency of payment management and supports the company's financial stability. However, the research is limited by the use of data from a single region, the absence of external factors in the model, and the high computational requirements. For broader implementation, further research should include data from other regions, consider external factors, and optimize computational resource usage.

Graph Regularized Probabilistic Latent Semantic Analysis for Topic Analysis Using Social Media Data

2025-01-08T20:05:25+00:00

In today's digital era, social media data provides valuable insights into public opinion. This study implements the Graph Regularized Probabilistic Latent Semantic Analysis (GPLSA) method to analyze topics from social media data surrounding the 2024 Indonesian Presidential Election (Pemilu), as well as to evaluate the efficiency of the Probabilistic Latent Semantic Analysis (PLSA) algorithm. The research stages include collecting social media data on presidential debates and elections, text pre-processing, and applying the GPLSA method to identify main topics. The analysis results show that PLSA without graph achieved a topic coherence score of 0.653, indicating good consistency, while GPLSA decreased to 0.5, suggesting that the addition of graph regularization did not significantly enhance coherence. Additionally, PLSA without graph achieved a perplexity score of 12.138, indicating good predictive capability, while GPLSA increased to 12.511, showing that graph regularization did not improve the prediction of new words. PLSA without graph also produced topics relevant to election issues, while GPLSA generated topics influenced by graph regularization, though without significant improvement in topic quality. Sentiment analysis of social media posts provides insights into public responses to debates and election issues. Validation of the GPLSA model ensures relevant topic representation. This research contributes to the development of text analysis methods and offers valuable information for elections and democratic participation. These results can be utilized by stakeholders to make more strategic and informed decisions.

Capability-Based API Gateway Technology Selection Analysis for Banking Cybersecurity Solution Using AHP Method

2024-12-14T19:04:55+00:00

The growing reliance on APIs in the banking sector, driven by digital transformation, necessitates robust API Gateways that balance performance with strong security measures to address risks like API abuse, man-in-the-middle attacks, and data scraping, while ensuring compliance with regulations such as PCI-DSS, GDPR, and OJK standards. This study bridges the gap in technical guidance by developing a comprehensive evaluation framework using the Analytic Hierarchy Process (AHP) to determine the most suitable API Gateway for banking. The findings identify Apigee as the optimal choice, scoring 1.4277 for its superior authentication, traffic encryption, threat detection, deployment flexibility, cloud integration, and API management. IBM API Connect, scoring 0.6186, is a strong alternative with excellent security and management features but limited scalability and deployment flexibility. Kong and Axway API Gateway follow with scores of 0.4215 and 0.4627, excelling in deployment and integration but lacking critical security features for banking. This research emphasizes the strategic importance of selecting the right API Gateway to bolster cybersecurity and API management in banking, recommending Apigee as the primary solution and IBM API Connect for complex IT infrastructures. It also contributes to the literature by providing a structured, quantitative approach to API Gateway selection and suggests future research exploring AI integration, advanced analytics, and cost-benefit analyses for informed decision-making in the financial sector.

Comparative Analysis of Random Forest and SVM Performance in Asthma Prediction

2024-12-24T16:42:21+00:00

This study evaluates the performance of Random Forest (RF) and Support Vector Machine (SVM) algorithms in predicting asthma risk to identify the most suitable method for medical datasets. Key metrics include training time, testing time, forecasting time, error rate, and accuracy. The datasets involve attributes such as age and clinical factors, analyzed in three stages: training, testing, and forecasting.

During training, SVM demonstrated faster processing, requiring a maximum of 0.5323 seconds compared to RF's 3.7736 seconds on 90% training data. In testing, RF excelled in speed, achieving a minimum testing time of 0.0215 seconds, while SVM required 0.0984 seconds on 10% test data. However, SVM consistently outperformed RF in error rate, with a minimum error rate of 19.17%, compared to RF's 25.10%.

Assessment Clusterization Teacher Performance with K-Means Algorithm Clustering and Agglomerative Hierarchical Clustering (AHC)

2024-10-28T17:45:20+00:00

Research This aims to do clustering evaluation teacher performance with the application of the K-means clustering algorithm and agglomerative hierarchical clustering (AHC). Background study This is based on needs to increase quality teaching through analysis and evaluation and better teacher performance. The methods applied involving assessment data collection performance from teachers in the environment education local, processed using a second algorithm The results of the research show that the silhouette score value for K-means reached 0.364, while AHC produced a value 0.343. With Thus, K-means is proven more effective in grouping assessment data and teacher performance compared to AHC. The conclusion of the study This confirms the importance of implementation of the K-means algorithm to get more insight into good evaluation teacher performance. Author Ready to do repairs or revisions to the manuscript. This is in accordance with comments and suggestions from the reviewer as a condition beginning. For processing more, carry on.

Fuzzy C-Means Algorithm for Grouping Students Based on Preferences and Academic Potential

2025-01-20T04:15:44+00:00

Personalized education is increasingly becoming a necessity in the modern era to ensure that students get a learning experience that is relevant to their interests and academic potential. This study aims to group students into three main clusters, namely Science, Arts, and Business, using the Fuzzy C-Means (FCM) algorithm. The FCM algorithm was chosen because of its flexibility in handling multidimensional data and allows students to have degrees of membership in more than one cluster, reflecting the multidisciplinary nature of their preferences. The research dataset consists of data on students' interests in fields of study (Science, Arts, Business) and academic grades in related subjects. The clustering results show that: The Business cluster includes 59 students (46.9%), reflecting the dominance of interests in economics, global trend analysis, and business organization activities. Artcluster consists of 39 students (30.0%), who show a preference for visual arts, art portfolio development, and involvement in community design. Science cluster has 30 students (23.1%) with interests in biology, science experiments, and biotechnology. Evaluation using Davies Bouldin Index (DBI) yields a value of 0.78, indicating good cluster quality. In addition, manual validation from teachers shows that more than 85% of students in each cluster fit the grouping based on direct observation. This study makes a significant contribution to the development of data-driven academic recommendation systems, enabling educational institutions to design learning programs that are more adaptive, relevant, and in accordance with student needs.

Improving Tesseract OCR Accuracy Using SymSpell Algorithm on Passport Data

2025-01-05T16:48:22+00:00

Optical Character Recognition (OCR) is a technology used to recognize text from images or digital documents, such as passports. One popular OCR tool is Tesseract as it offers high accuracy. However, OCR accuracy is often affected by various factors, including image noise and/or non-text elements. This article discusses the application of the SymSpell algorithm for post processing to improve OCR accuracy on standard Indonesian passports. OCR will be focused on the Visual Inspection Zone, specifically the Place of Birth and Issuing Office values. Unlike the Machine Readable Zone which is composed of individual codes and a clear background, the Visual Inspection Zone often experiences OCR errors due to holograms blocking the text and spaced layouts. SymSpell is an edit distance based spelling correction algorithm designed to process data quickly and efficiently, even on very huge datasets. In this study, SymSpell is used to detect and correct errors in OCR results that are compared to a corpus word list. Experimental results with 10 tested scans and passport photos showed that the integration of SymSpell with the Research and Development methodology was able to improve the OCR accuracy rate by 21,43% for certain Place of Birth and Issuing Office data from the Visual Inspection Zone. With this approach, OCR systems can provide more reliable results for practical applications.

Assessment Of IDW And ANN On Daily Rainfall Data Imputation in Semarang Central Java

2025-01-13T18:25:29+00:00

Rainfall plays a critical role in the global water and energy cycle, influencing surface water availability and recharge processes both spatially and temporally. Traditional rainfall data collection using ombrometers provides accurate live data, but often faces the challenge of missing data due to equipment failure or transmission, especially in agencies such as BMKG. This problem of missing data greatly impacts hydrological analysis and requires an effective data recovery process through imputation. This study aims to assess the accuracy of rainfall data imputation techniques using the Inverse Distance Weighting (IDW) and Artificial Neural Network (ANN) methods. In this study, we utilize data from 31 observation stations in Semarang City for more than three decades. The findings show that the spatial distribution of rainfall is variable and exhibits a cyclic pattern despite fluctuations. The ANN model performed very well in overcoming missing data, especially in the dry season with an RMSE of 0.9489 and a coefficient of determination (R²) of 0.9926. By demonstrating the superiority of the ANN model in accurately predicting rainfall, this study offers an effective approach to improve the quality of BMKG climate data. This is expected to support disaster mitigation decisions and sustainable development planning. This approach demonstrates that the selection of an appropriate method is critical for accurate and reliable analysis of rainfall time series data. In addition to making an academic contribution, these results also provide an alternative imputation method for various time series.

Predicting Prospective Student Interests Using the C4.5 Algorithm and Naive Bayes

2025-01-12T19:10:07+00:00

Students are individuals pursuing higher education at a university with the goal of enhancing their knowledge, skills, and character to succeed in the professional world and contribute to society. The purpose of this study is to analyze the factors that influence prospective students' interest in continuing their education using the C4.5 Algorithm and the Naïve Bayes Method. The importance of understanding prospective students' interest patterns is expected to help universities formulate more effective strategies. The purpose of this study is to determine how well the two methods classify data and understand the factors that most influence prospective students' decisions. The C4.5 Algorithm is known to be effective in building decision trees that are easy to interpret, while the Naïve Bayes Method has the advantage of handling datasets with independent attributes. This study uses the stages of data selection, data pre-processing, algorithm application, and model evaluation. The classification results obtained from the C4.5 Algorithm show that 132 data are included in the interest category and 8 data are not interested, while the Naïve Bayes Method produces 131 data of interest and 9 data are not interested. In conclusion, both methods have good accuracy levels, but the Naïve Bayes Method shows superiority in Recall value, while the C4.5 Algorithm excels in interpretation of results and clarity of classification patterns.

Comparative Analysis of Express and Hono Framework Performance in Simple Registration Application

2024-12-18T19:01:34+00:00

This research evaluates the performance of two Node.js frameworks, Express and Hono, in developing a simple registration application. This application serves as a backend to store user registration data into a PostgreSQL database using the pg client of the node package manager (npm). The purpose of this performance comparison is to identify the framework that is superior in executing 1 million requests in this scenario. The analysis shows that Express has an average execution time of 26.85% faster than Hono. However, it is inversely proportional to the resource usage, where Hono shows better efficiency with lower CPU and memory usage of 29.29% and 19.97%. These findings provide important insights for developers in choosing a suitable framework based on performance and resource efficiency requirements.

Sentiment Analysis On Indonesian Tweets about the 2024 Election

2025-01-18T06:30:52+00:00

This study investigates public sentiment on Indonesian Twitter regarding the 2024 General Election, employing machine learning and deep learning techniques, including Naïve Bayes, Support Vector Machines (SVM), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU). The dataset was collected using a Tweet Harvest method with the keyword "Pemilu" and underwent preprocessing steps such as case folding, removal of symbols and URLs, stopword elimination, and tokenization to ensure data quality. Performance evaluation metrics, including accuracy, precision, recall, and F1-score, were applied to assess the models' effectiveness. Naïve Bayes achieved the highest accuracy of 64%, followed by SVM at 63%, LSTM at 60%, and GRU at 57%. The findings indicate that traditional models like Naïve Bayes and SVM perform effectively on smaller datasets with structured features, while deep learning models excel in capturing complex sequential dependencies. However, deep learning methods exhibited overfitting tendencies, indicating the need for better regularization and optimization techniques. Furthermore, it emphasizes the potential of integrating traditional algorithms with advanced methods to enhance sentiment classification accuracy and generalizability across diverse datasets.

Optimization of Player Experience and Enemy AI using A* Algorithm in Game

2025-01-13T18:17:35+00:00

The gaming industry is rapidly evolving, where engaging and challenging gameplay has become a key factor in a game's success. Effective enemy intelligence can enhance challenges and enrich the player experience. This study aims to improve the player experience and enemy intelligence in the game Galang the EcoRescue through the implementation of the A-star (A*) algorithm. A* is a pathfinding algorithm that uses distance estimation to find the shortest path to a target by utilizing a heuristic function. This game was developed using the Unity Engine, with the implementation of the A* algorithm to determine enemy movements and adapt their behavior according to the game’s situation. Testing was conducted to ensure improvements in both the player experience and enemy intelligence. The results of the study show that the A* algorithm successfully enhanced enemy intelligence by creating more realistic and adaptive movements in response to the player, ultimately providing more dynamic challenges and improving overall gameplay quality. This study utilized the Game Development Life Cycle (GDLC) method, covering the stages of initiation, pre-production, production, testing, beta, and release. The A* algorithm has proven to optimize the player experience in Galang the EcoRescue.

Effectiveness of Bi-GRU and FastText in Sentiment Analysis of Shopee App Reviews

2025-01-19T04:20:38+00:00

E-commerce is proof of evolution in the economic field due to its flexibility to shop for various necessities of life anytime and anywhere. Shopee is one of the e-commerce platforms in demand by people from varied circles in Indonesia. Multiple reviews are shed publicly by Shopee users on the Google Play Store regarding shopping experiences, which can be positive or negative. This condition affects the decision of other users to shop at Shopee, thus impacting the increase or decrease in profits from Shopee itself. Therefore, user sentiment analysis is needed as a form of effort to maintain user trust in Shopee. This research aims to build a system to classify the sentiment of Shopee application users through reviews in the Google Play Store by utilizing the Bidirectional Gated Recurrent Unit (Bi-GRU) deep learning model. The dataset contains 9,716 reviews, including 3,937 positive and 5,779 negative sentiments. Several test scenarios were conducted to achieve the highest peak of performance, utilizing TF-IDF feature extraction, FastText feature expansion, and optimization using the Cuckoo Search Algorithm. Additionally, SMOTE resampling was utilized to correct the dataset’s uneven distribution. The combined test scenarios mentioned significantly improved the accuracy by 1.03% and F1-Score by 1.04% from the baseline, with the highest accuracy reaching 90.48% and the highest F1-Score of 90.16%.

Comparison of C4.5 & Random Forest Based on AdaBoost For Determining Loan Eligibility Customer Funds

2025-01-21T20:09:50+00:00

This research discusses the comparison between two data mining algorithms, namely Decision Tree C4.5 and Random Forest based on AdaBoost, in determining the creditworthiness of customer funds. The main objective of this research is to evaluate and compare the performance of the two algorithms in predicting loan eligibility based on customer data. Algorithm performance is measured using accuracy, precision, recall, and misclassification error metrics. The research results show that the AdaBoost-based Random Forest is superior with an accuracy of 78.86%, recall of 98.75%, and the lowest misclassification error of 21.14%. Meanwhile, Decision Tree C4.5 provides lower performance than AdaBoost-based Random Forest. This research recommends further exploration of other algorithms, such as Support Vector Machine (SVM) and Neural Networks, to obtain more optimal results in determining customer loan eligibility.

Decision-Making Framework Using MARCOS for Evaluating Sealing Machines in Small and Medium Enterprises

2024-12-24T16:50:17+00:00

In the era of globalization, Micro, Small, and Medium Enterprises (MSMEs) hold a vital position in Indonesia's economy, contributing significantly to GDP and employment. Despite their importance, MSMEs need help in selecting appropriate sealer machines, which affects production efficiency and product quality. There are six different kinds of sealer machines that are looked at in this study. They are manual, vertical, continuous, horizontal semi-automatic, impulse, and vacuum. The MARCOS method is used to find the best option. Results indicate that the Impulse Sealer Machine (A5) is the most suitable, with a Ki value of 1.7, followed by the Continuous Sealer Machine (A3), with a Ki of 1.63. Machines such as Manual (A1), Vertical (A2), and Vacuum (A6) scored 1.6, while the Horizontal Semi-Automatic Sealer Machine (A4) ranked lowest at 1.36. These findings provide MSMEs with practical guidance for selecting sealer machines that enhance production efficiency and competitiveness in the global market while also contributing to the development of packaging technology research.

A Comparative Analysis of Clustering Algorithms for Expedia’s Travel Dataset

2024-12-24T16:40:48+00:00

The effective segmentation of travel data is crucial for deriving actionable insights in the tourism and hospitality sectors. This study conducts a comprehensive evaluation of four clustering algorithms Agglomerative Clustering, DBSCAN, Gaussian Mixture Models (GMM), and KMeans on a travel dataset, using three widely recognized metrics: Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Score. The dataset was preprocessed through standardization and dimensionality reduction via Principal Component Analysis (PCA) to facilitate visualization and ensure computational efficiency. The results highlight significant differences in the performance of these algorithms. Agglomerative Clustering achieved the highest Silhouette Score, indicating superior cluster cohesion and separation, while KMeans recorded the highest Calinski-Harabasz Score, demonstrating strong inter-cluster variance. In contrast, DBSCAN performed poorly, producing low scores across all metrics, attributed to sensitivity to parameter selection and density irregularities in the dataset. Gaussian Mixture Models exhibited moderate performance but struggled with overlapping clusters due to limitations in modeling non-Gaussian data distributions. Visualization of clustering results confirmed these findings, revealing compact clusters for Agglomerative and KMeans, while DBSCAN and GMM showed less defined structures. This study underscores the importance of selecting clustering algorithms based on dataset characteristics and analysis objectives

ISO 27001 As Information Security Solution In Society 5.0 Era: Systematic Literature Review

2025-01-13T18:22:21+00:00

In the era of Society 5.0, information security is an important issue along with the increasing use of technologies such as the Internet of Things (IoT), Artificial Intelligence (AI), and big data. ISO 27001 acts as a globally recognized standard framework for managing information security. The ISO 27001 standard provides a systematic framework for identifying, assessing, and managing information security risks so as to ensure the integrity, confidentiality, and availability of data in an organization. This research aims to evaluate the implementation of ISO 27001 as an information security solution in the Society 5.0 era through a systematic literature review. Using the Systematic Literature Review (SLR) method, this research collects and analyzes relevant literature to identify benefits, challenges, and recommendations related to the application of ISO 27001 in an era of increasingly integrated technology. The results showed that the implementation of ISO 27001 in the Society 5.0 era proved to make a significant contribution in improving organizational information security. This is done through a PDCA (Plan-Do-Check-Act) approach that integrates information security policies into business processes, strengthens risk management, technology infrastructure, and human resource competencies. In conclusion, the implementation of ISO 27001 in the Society 5.0 era not only improves information security, but also supports the achievement of operational efficiency and organizational sustainability amid rapid technological developments.

A Comparative Study of Ensemble Learning and Neural Networks for the Heart Disease Prediction

2024-12-25T07:06:13+00:00

Heart disease continues to be a leading global cause of death, making the development of predictive models for early diagnosis a critical task. This study investigates the performance of various machine learning and deep learning models for heart disease prediction using a structured dataset of 918 observations and 11 features. The analysis includes ensemble methods like Random Forest, Gradient Boosting, and XGBoost, as well as neural networks such as Multi-Layer Perceptrons (MLPs) and Convolutional Neural Networks (CNNs). Traditional classifiers, including Support Vector Machines (SVM) and Logistic Regression, are also considered for benchmarking. The dataset was preprocessed using label encoding, standardization, and the Synthetic Minority Oversampling Technique (SMOTE) to address class imbalance and ensure data consistency. Model evaluation was conducted using key metrics such as precision, recall, F1-score, and ROC-AUC. The results demonstrated that ensemble methods, particularly Random Forest (ROC-AUC: 0.9313) and Gradient Boosting (ROC-AUC: 0.9279), consistently delivered superior performance. Among neural networks, MLPs showed promising results (ROC-AUC: 0.9232), outperforming CNNs, which were less effective in handling tabular data. Meanwhile, TabNet was found to be unsuitable for this dataset, as it significantly underperformed across all metrics. This research highlights the effectiveness of ensemble methods and MLPs in heart disease prediction and the importance of proper preprocessing techniques. Future work could focus on integrating hybrid models or advanced optimization techniques to further enhance predictive accuracy in clinical settings.

Comparison of K-Nearest Neighbor, Naive Bayes, Random Forest Algorithms for Obesity Prediction

2025-01-19T05:00:10+00:00

Obesity is a global health problem that continues to increase and has serious impacts on physical and mental health. This research aims to predict a person's obesity status based on certain attributes using the K-Nearest Neighbor (KNN), Naive Bayes, and Random Forest algorithms. The dataset used was taken from the Kaggle platform with 2,111 data and 16 attributes, including gender, age, weight, height, frequency of consumption of high-calorie foods, physical activity, and water and vegetable consumption patterns. The research process follows the data mining stages, including business understanding, data understanding, data preparation, modeling, evaluation, and documentation. Experiments were carried out using RapidMiner with a cross-validation technique using 10 folds to measure overall model performance. The research results show that the Random Forest algorithm performs best in predicting obesity status compared to K-NN and Naive Bayes. Model evaluation using accuracy, precision, recall, and F1-score metrics shows significant results in distinguishing obesity categories. It is hoped that this research can contribute to the development of a machine learning-based health prediction system that can be used to support decision-making in the prevention and management of obesity.

Comparative Analysis of Homogeneous and Heterogeneous Ensembles for Diabetes Classification Optimization

2025-01-28T05:10:06+00:00

Diabetes mellitus is a chronic disease with an increasing prevalence worldwide, including in Indonesia, reaching 11.7% by 2023. Early prediction of this disease is essential for more effective management. This study aims to develop a diabetes mellitus prediction model using an ensemble learning approach, including homogeneous (boosting and bagging) and heterogeneous (stacking and blending) techniques. In this study, the boosting algorithm using AdaBoost with Random Forest as the base estimator showed the highest accuracy of 98%, with balanced precision and recall. The bagging technique, which also uses Random Forest as the base estimator, achieved 97% accuracy, although slightly lower than boosting. The stacking technique, which combines XGBoost, Gradient Boosting, and Random Forest as base learners, with Random Forest as the meta-model, yields similar accuracy of 98%, but with lower prediction error, demonstrating its ability to cope with more complex data. Blending, which uses a similar approach but with training on the entire dataset, gave 98% accuracy with shorter processing time and more efficient memory usage than stacking.

The Efficiency of Machine Learning Techniques in Strengthening Defenses Against DDoS Attacks, Such as Random Forest, Logistic Regression, and Neural Networks

2025-01-21T04:39:22+00:00

Distributed Denial of Service (DDoS) attacks are one of the most common cybersecurity concerns brought on by the quick development of digital technology. By flooding servers with too many requests, these assaults interfere with online services, highlighting the necessity of strong detection systems. Using the well-known CIC-DDoS2019 dataset, this study explores the use of machine learning algorithms—Random Forest (RF), Logistic Regression (LR), and Neural Networks (NN)—to improve DDoS assault detection. A comprehensive preprocessing procedure that comprised feature selection, normalization, and duplication removal was applied to dataset in order to ensuring optimal algorithm performance. With an accuracy of 97% on the entire test dataset and 99.13% on the training and validation datasets, RF showed exceptional performance. While NN successfully managed intricate data patterns, attaining an accuracy of roughly 94%, LR demonstrated impressive results with an accuracy of 98.65%. Because of its ensemble method, which minimizes overfitting and improves model generalization, the RF algorithm performed better than the others. This study highlights how machine learning may be used to solve practical cybersecurity issues by offering insightful information about how to optimize algorithms for real-time DDoS detection. The results improve the stability and resilience of digital infrastructures by aiding in the creation of effective intrusion detection systems. Future research can explore integrating advanced neural network architectures and hybrid methods to further improve detection rates and adaptability to evolving cyber threats.

The Enterprise Architecture for Enhanced Mutual Fund Service Integration in Digital Channel in the Banking Industry

2025-02-06T19:39:27+00:00

The investment awareness among the Indonesian populace is rising, particularly during the Covid-19 pandemic. Mutual funds are the most prevalent investment vehicle due to their accessibility, requiring minimal cash, and offering competitive returns, as they are managed by seasoned investment professionals. This is due to the relative ease of investing in this instrument with accessible capital and competitive returns, as fund management is conducted by seasoned investment managers. Conversely, the rapid advancement and proliferation of technology present new challenges for firms in the financial sector, particularly in banking and financial technology, as they strive to innovate and provide convenient services that cater to the needs of customers and investors. This research seeks to develop Enterprise Architecture for the integration of Mutual Fund services into digital channels via mobile banking, utilizing the TOGAF framework as the foundational design approach, supplemented by SWOT analysis to assess the strategic position and market potential. The research employed a qualitative methodology utilizing the ArchiMate program to illustrate diagrams in Enterprise Architecture (EA) across the public banking sector. The research indicated that architectural design can facilitate improved data access and enhance adaptability to technology and market advancements, hence removing inefficiencies and streamlining the review process. In conclusion, the application of EA in the incorporation of mutual fund services into the digital banking platform will optimize the company's performance processes to attain objectives, while also enhancing agility in evaluating technological risks to facilitate the monitoring, metrics, and analysis of information technology, thereby ensuring the achievement of business goals.