Analisis Sentimen Pengguna terhadap Aplikasi Lalamove dengan Perbandingan Algoritma Support Vector Machine dan Naive Bayes

Abstract

The development of digital technology has brought significant changes to the logistics and transportation sector. On-demand delivery applications such as Lalamove are a solution for users who need fast and efficient services. However, the existence of various user reviews both positive and negative indicates a difference in experience that needs to be analyzed. This study aims to evaluate user perceptions of the Lalamove application by comparing the effectiveness of Support Vector Machine (SVM) and Naive Bayes algorithms in sentiment classification. The data used in this study includes 10,000 user reviews obtained through scraping techniques from the Google Play Store. After going through the data preprocessing stage, the analysis is performed using TF-IDF method as feature extraction and the model performance evaluation is performed based on accuracy, precision, recall, and F1-score metrics. Sentiment classification in this study was performed in two categories, namely positive and negative (binary) sentiment, without considering the neutral category. The results show that the SVM algorithm has a high accuracy of 87% compared to Naive Bayes which only reaches 83%. This research provides an understanding for application developers in improving service quality based on user sentiment analysis.