<![CDATA[This research compares the accuracy of the K-Nearest Neighbor (KNN) and Naive Bayes methods in classifying user sentiment towards the DANA e-wallet application using Term Frequency-Inverse Document Frequency (TF-IDF) feature extraction. User review data was collected through web scraping techniques and labeled by linguists and lexicon models. After undergoing pre-processing steps such as case folding, cleaning, tokenizing, stopword removal, and stemming, the data was classified using the KNN and Naive Bayes methods. The research results indicate that data labeling by linguists significantly improves the accuracy of both classification methods. Additionally, using TF-IDF as a word weighting method proves effective in enhancing the performance of sentiment classification models. Sentiment analysis of user reviews of the DANA application reveals various complaints and issues faced by users, providing information that can be used to improve the features and services offered, thereby increasing user satisfaction. This research also provides a comparison between the KNN and Naive Bayes methods, which can serve as a reference for other researchers in selecting appropriate methods for sentiment analysis on similar datasets.]]>
Copyrights © 2024
