Sentiment Analysis on Service Provider Reviews Using the C4.5 Algorithm

Abstract

Sentiment analysis is essential for understanding and processing textual data to derive meaningful insights. PT. XYZ, a company operating in the GSM cellular telecommunications sector, faces a challenge due to the lack of a specific application for analyzing visitor reviews on their services. This gap impedes their ability to gain detailed insights into consumer feedback, hindering efforts to improve service quality. This research addresses this issue by developing an application that utilizes the C4.5 algorithm to analyze PT. XYZ's reviews. The study uses 140 consumer reviews collected from Google Maps. The C4.5 algorithm, which creates decision trees to find relationships among variables, is employed for classifying and predicting service review sentiments. The research involves several stages: data crawling, text preprocessing, term frequency (TF) calculation, and applying the C4.5 algorithm for classification.The results demonstrate the effectiveness of this approach. With 126 training data samples and 14 test samples, the model achieved an accuracy of 78.57%, precision of 83.33%, and recall of 90.91%. These findings indicate that increasing the amount of training data enhances pattern recognition and accuracy. The study successfully meets its objectives, proving that sentiment analysis using the C4.5 algorithm can effectively predict service review sentiments and aid in improving service quality.