Analisis Sentimen untuk Perbaikan Kualitas Layanan M-Paspor Imigrasi

Abstract

The purpose of this research is to identify and analyze the implementation results of the Support Vector Machine (SVM) and Naive Bayes Classifier (NBC) methods in a case study of public sentiment analysis toward the M-Paspor application on Google Play Store. Based on the findings, this research aims to improve the quality of services provided by the M-Paspor application. This study employs descriptive research that systematically, factually, and accurately describes or explains a specific phenomenon, condition, or object. It specifically seeks to illustrate and gather facts about user sentiment regarding the M-Paspor application based on reviews on the Google Play Store platform. The data used in this study are secondary data collected from M-Paspor application reviews from June 2022 to June 2024 using web scraping techniques. Based on the labeling process using Python on the Google Collaboratory platform, it was found that 12.32% of the reviews were positive, while 87.67% were negative. The sentiment analysis using the Support Vector Machine (SVM) and Naïve Bayes Classifier (NBC) algorithms showed that the best accuracy was achieved by the SVM method, with an accuracy of 98%, precision of 97%, recall of 98%, and an F-measure of 97% In contrast, the NBC method obtained an accuracy of 87%, precision of 88%, recall of 84%, and an F-measure of 86%. These results indicate that the Support Vector Machine (SVM) algorithm performed better in sentiment classification for the M-Paspor application. Based on the identification of service quality variables in the negative sentiment class, the most common reason for negative comments from users was the reliability factor, which scored the highest percentage at 43.5%. This was followed by the responsiveness factor at 27.6%, the empathy factor at 11.2%, the tangibility factor at 8.9%, and the assurance factor at 8.7%.