Automated Vader Lexicon Labeling on Sentiment Analysis Against The E-Wallet

Abstract

In Indonesia, e-wallets have gained immense popularity due to their user-friendly interfaces and attractive features. Despite their convenience, user opinions about e-wallet applications remain polarized, with sentiments ranging from highly positive to critically negative. This study seeks to analyze these diverse user sentiments by leveraging the VADER Lexicon model, a powerful tool for sentiment analysis. The Naïve Bayes Classifier, a well-established probabilistic model renowned for its efficacy in text-based classification tasks, is employed to categorize user reviews. The sentiment analysis yielded promising results, with the model achieving an impressive accuracy rate of 92.02%. Additionally, the precision of the model, indicating the ratio of correctly predicted positive sentiments to all predicted positive sentiments, stood at 83.23%. The recall, representing the ratio of correctly predicted positive sentiments to all actual positive sentiments, was recorded at 86.80%. These metrics underscore the model's robustness in accurately classifying user sentiments. The insights gained from this analysis provide a deeper understanding of user perspectives, aiding in the evaluation and enhancement of e-wallet applications.