Comment Intention Classification in TikTok Fashion Affiliate Marketing Based on Attention-Weighted Word2Vec Representation
DOI:
10.33395/sinkron.v10i4.16843Keywords:
Intention Classification, Logistic Regression, Norm-Based Softmax Attention, Support Vector Machine, TikTok Affiliate, Word2VecAbstract
TikTok Fashion Affiliate generates short, informal user comments with a variety of communicative intentions, necessitating a text representation capable of preserving key information within the comments. This study aims to evaluate Norm-Based Softmax Weighting as a non-learned weighting mechanism and compare it with Mean Pooling for classifying the intent of Indonesian-language TikTok Fashion Affiliate comments. The comment dataset was independently annotated by three annotators with a Krippendorff’s Alpha of 0.7950 and classified into three classes: Question, Request, and Opinion. Word2Vec Skip-gram with a dimension of 100 was used to generate comment representations, which were then compared using Mean Pooling and Norm-Based Softmax Weighting in Support Vector Machine (SVM) and Logistic Regression models. Hyperparameters were selected via Grid Search using Macro F1 as the optimization criterion. To prevent data leakage, Word2Vec was trained only on the training data and retrained on each fold. Evaluation was performed using the test data, 5-Fold Stratified Cross-Validation, and Leave-One-Video-Out. The results showed that SVM with Norm-Based Softmax Weighting achieved a Macro F1 score of 0.7995, slightly higher than Mean Pooling’s 0.7913. In contrast, for Logistic Regression, Mean Pooling achieved a Macro F1 score of 0.7846, which was higher than the 0.7768 achieved by Norm-Based Softmax Weighting. However, ablation analysis and statistical testing did not reveal any significant differences. Leave-One-Video-Out validation revealed performance variations across videos and a decline in performance compared to random data splitting. Thus, Norm-Based Softmax Weighting is a competitive non-learned weighting alternative, but its effectiveness depends on the classifier and data characteristics.
Downloads
References
Andriyani, W., Astuti, Y. and Wisesa, B.A. (2024) “Analisis Sentimen pada Ulasan Produk dengan SVM dan Word2Vec Sentiment Analysis on Product Reviews with SVM and Word2Vec,” (1), pp. 173–185. Available at: https://doi.org/10.26798/jiko.v8i1.1498.
Asramanggala, M.S., Prasetyowati, S.S. and Sibaroni, Y. (2023) “Optimal Number Data Trains in Hoax News Detection of Indonesian using SVM and Word2Vec,” 5(1), pp. 21–28. Available at: https://doi.org/10.47065/bits.v5i1.3516.
Aubaid, A.M., Mishra, Alok and Mishra, Atul (2024) “Machine learning and rule ‑ based embedding techniques for classifying text documents,” International Journal of System Assurance Engineering and Management, 15(12), pp. 5637–5652. Available at: https://doi.org/10.1007/s13198-024-02555-w.
Ayuningtyas, P. and Tantyoko, H. (2024) “Perbandingan Metode Word2vec Model Skipgram pada Ulasan Aplikasi Linkaja menggunakan Algoritma Bidirectional LSTM dan Support Vector Machine Comparison of the Word2vec Skipgram Model Method Linkaja Application Review using Bidirectional LSTM Algorithm and Support Vector Machine,” 12(1), pp. 189–196. Available at: https://doi.org/10.26418/justin.v12i1.72530.
Harahap, Y.A. and Sazali, H. (2024) “Digital Marketing Tiktok Affiliate pada Pemasaran Produk Fashion,” 6(5), pp. 1733–1743.
Hashmi, E. et al. (2025) “Multi-domain text classification via linguistic and semantic feature integration.”
Hendrawan, I.R., Utami, E. and Hartanto, A.D. (2022) “Analisis Perbandingan Metode Tf-Idf dan Word2vec pada Klasifikasi Teks Sentimen Masyarakat Terhadap Produk Lokal di Indonesia,” pp. 497–503.
Kaibassova, D. (2025) “Document analysis via combined vectorization and machine learning approaches,” 8(4), pp. 2195–2204. Available at: https://doi.org/10.53894/ijirss.v8i4.8356.
Li, J. et al. (2025) An Improved Attention-Based LSTM Neural Network for Intelligent Anomaly Detection in Financial Statements, 2025 2nd International Conference on Artificial Intelligence, Digital Media Technology and Interaction Design (ICADI 2025), November 2830, 2025, Tianjin, China. Association for Computing Machinery. Available at: https://doi.org/10.1145/3795926.3796015.
Maulani, R., Ula, M. and Retno, S. (2026) “Klasifikasi teks berbahasa indonesia menggunakan word2vec dan k-nearest neighbor untuk identifikasi berita hoax 1) 1,2,3),” 11(1), pp. 122–129.
Media, J. and Budidarma, I. (2022) “Feature Expansion Using Word2vec for Hate Speech Detection on Indonesian Twitter with Classification Using SVM and Random,” 6(April), pp. 979–988. Available at: https://doi.org/10.30865/mib.v6i2.3855.
Muchori, J.G. and Maina, M.P. (2025) “Review Article A Systematic Literature Review on Application of Text Embedding Techniques in NLP,” 13(December), pp. 45–61.
Priyadarshini, J. (2025) “Enhanced Hybrid Deep Learning Model With Improved Self-Attention Mechanism for Legal Judgment Prediction,” IEEE Access, 13(August), pp. 139868–139882. Available at: https://doi.org/10.1109/ACCESS.2025.3596180.
Pujianto, D. and Asia, U.M. (2025) “Journal of Computer Networks , Architecture and High Performance Computing APPLICATION OF TRANSFORMER MODEL AND WORD EMBEDDING IN SENTIMENT ANALISYS OF INDONESIAN E- Journal of Computer Networks , Architecture and High Performance Computing,” 7(3), pp. 720–732.
Riana, D. et al. (2025) “WORD2VEC OPTIMALIZATION USING TRANSFER LEARNING IN,” 11(2), pp. 382–390. Available at: https://doi.org/10.33480/jitk.v11i2.6051.dataset.
Rifaldy, F., Sibaroni, Y. and Prasetiyowati, S.S. (2025) “Effectiveness of word2vec and tf-idf in sentiment classification on online investment platforms using support vector machine 1.,” 10(2), pp. 863–874.
Rinaldi, M.K. et al. (2026) “Application of Two-Stream Late Fusion on EfficientNetV2 based on Transfer Learning to classify AI-generated paintings,” 10(2), pp. 976–990.
Rizky, F.M. and Lhaksmana, K.M. (2023) “Twitter Sentiment Analysis of Kanjuruhan Disaster using Word2Vec and Support Vector Machine,” 5(1), pp. 219–227. Available at: https://doi.org/10.47065/bits.v5i1.3612.
Sahbuddin, M. and Agustian, S. (2022) “JITE ( Journal of Informatics and Telecommunication Engineering ) Support Vector Machine Method with Word2vec for Covid-19 Vaccine Sentiment Classification on Twitter,” 6(July), pp. 288–297.
Saputra, J. et al. (2025) “Analisis Performa Naive Bayes dan SVM terhadap Sentimen Teks Media Sosial dengan Word2Vec dan SMOTE,” 10(1), pp. 143–155.
Science, C. and Fauzi, M.A. (2019) “Word2Vec model for sentiment analysis of product reviews in Indonesian language,” 9(1), pp. 525–530. Available at: https://doi.org/10.11591/ijece.v9i1.pp525-530.
Suharman, A. et al. (2025) “Analisis Sentimen Pengguna Aplikasi Livin ’ by Mandiri Menggunakan Metode Support Vector Machine ( SVM ) dengan Ekstraksi Fitur TF-IDF dan Word2Vec User Sentiment Analysis of the Livin ’ by Mandiri Application Using the Support Vector Machine ( SVM ) Method with TF-IDF and Word2Vec Feature Extraction,” 5(8), pp. 2201–2212.
Wicaksana, H.S. and Huda, K. (2025) “Penerapan Word2Vec dan SVM dengan Hyperparameter Tuning untuk Deteksi Phishing,” 12(3), pp. 361–371. Available at: https://doi.org/10.30865/jurikom.v12i3.8729.
Yousefi, R. et al. (2026) “Enhancing logistic regression classification : insights from simulation and real-world applications through ranked set sampling,” pp. 1–28.
Zhao, W. et al. (2022) “WTL-CNN : a news text classification method of convolutional neural network based on weighted word embedding,” 0091. Available at: https://doi.org/10.1080/09540091.2022.2117274.
Downloads
|
How to Cite
Issue
Section
License
Copyright (c) 2026 Rahmawati Septiani, Masy Ari Ulinuha, Nur Cahyo Hendro
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Muhammad Khoiruddin harahap
Google Scholar 
Moraref
PKP Index
Garuda Kemdikbud
Indonesia OneSearch
OCLC Worldcat
Dimensions
Index Copernicus
Scilit
