Comparative Analysis of the Performance of K-Nearest Neighbor (K-NN) and Naive Bayes Algorithms on User Satisfaction Levels of the Tokopedia Application
Proceedings of The International Conference on Computer Science, Engineering, Social Science, and Multi-Disciplinary Studies · 02.11.2025 · 1 · 218-224
DOI: 10.64803/cessmuds.v1.37 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of The International Conference on Computer Science, Engineering, Social Science, and Multi-Disciplinary Studies, online erschienen am 02.11.2025. Verfasst von Muhammad Syahputra Novelan und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
The rapid advancement of digital technology has significantly influenced the development of e-commerce platforms in Indonesia, where Tokopedia stands out as one of the most popular and widely used online marketplaces. As user expectations continue to increase, understanding and measuring user satisfaction has become essential for ensuring service quality and maintaining customer loyalty. This study aims to perform a comparative analysis of the performance of two machine learning classification algorithms—K-Nearest Neighbor (K-NN) and Naive Bayes—in analyzing and predicting user satisfaction levels toward the Tokopedia application. The dataset used in this study was obtained from a combination of online reviews and structured survey responses from active Tokopedia users. The research methodology includes several stages: data collection, text preprocessing (tokenization, stop-word removal, and stemming), feature extraction using the Term Frequency–Inverse Document Frequency (TF-IDF) technique, and model implementation using the two algorithms. Both models were evaluated using key performance metrics such as accuracy, precision, recall, and F1-score. The experimental results indicate that the K-NN algorithm achieved superior performance compared to Naive Bayes, demonstrating higher accuracy and better consistency in classifying user sentiments into “satisfied” and “dissatisfied” categories. The K-NN model proved to be more effective in handling diverse and nonlinear data patterns derived from user-generated reviews. Meanwhile, Naive Bayes, although computationally efficient, showed limitations in processing complex text dependencies. The findings of this research highlight the importance of selecting appropriate machine learning algorithms for user satisfaction analysis. Furthermore, the study contributes to the broader understanding of sentiment-based evaluation models in e-commerce platforms and provides valuable insights for Tokopedia and similar companies in enhancing customer experience and service improvement strategies.
Zitieren
APA
Novelan, M. S., & Iqbal, M. (2025). Comparative Analysis of the Performance of K-Nearest Neighbor (K-NN) and Naive Bayes Algorithms on User Satisfaction Levels of the Tokopedia Application. Proceedings of The International Conference on Computer Science, Engineering, Social Science, and Multi-Disciplinary Studies, 1, 218-224. https://doi.org/10.64803/cessmuds.v1.37
Vancouver
Novelan MS, Iqbal M. Comparative Analysis of the Performance of K-Nearest Neighbor (K-NN) and Naive Bayes Algorithms on User Satisfaction Levels of the Tokopedia Application. Proceedings of The International Conference on Computer Science, Engineering, Social Science, and Multi-Disciplinary Studies. 2025;1:218-224. doi:10.64803/cessmuds.v1.37
BibTeX
@article{Novelan2025,
author = {Novelan, Muhammad Syahputra and Iqbal, Muhammad},
title = {Comparative Analysis of the Performance of K-Nearest Neighbor (K-NN) and Naive Bayes Algorithms on User Satisfaction Levels of the Tokopedia Application},
journal = {Proceedings of The International Conference on Computer Science, Engineering, Social Science, and Multi-Disciplinary Studies},
year = {2025},
volume = {1},
pages = {218--224},
doi = {10.64803/cessmuds.v1.37}
}
EUVIMED