<![CDATA[This study aims to analyze user sentiment towards the STRAVA app based on reviews obtained from the Google Play Store. The data used amounted to 6,082 reviews after going through text cleaning and pre-processing stages, including case folding, stopword removal, tokenizing, and stemming. Reviews with ratings of 4–5 are categorized as positive sentiment, while ratings of 1–2 are categorized as negative sentiment. The Naive Bayes algorithm was used as a classification model with a feature representation based on Term Frequency–Inverse Document Frequency (TF-IDF). The evaluation results showed that the model was able to classify sentiment with an accuracy of 90.7%, precision and recall for positive sentiment reached 94%, while for negative sentiment each was around 78%. These findings indicate that Naive Bayes with TF-IDF weighting is effective in analyzing sentiment of STRAVA app reviews. This study is expected to provide input for STRAVA developers in improving feature quality, user experience, and overall satisfaction.]]>
Copyrights © 2025
