Sentiment Analysis Of The Indrive Application Reviews On Google Play Store Using The Naive Bayes Method

Abstract

The development of application-based technology on Android smartphones has transformed daily activities, including the transportation sector. One popular online transportation application is inDrive, offering features such as fare negotiation and driver selection criteria. With over 100 million downloads and 8 million reviews on Google Play Store, sentiment analysis of user reviews is crucial for evaluating user satisfaction and providing insights for application developers. This study aims to analyze user sentiment on inDrive app reviews from the Google Play Store using the Naïve Bayes algorithm. The research adopts a quantitative approach with a dataset of 1,000 reviews collected through web scraping using Google Colab. The methodology includes sentiment labeling (positive and negative), data preprocessing (case folding, tokenizing, stopword removal, and stemming), weighting using Term Frequency-Inverse Document Frequency (TF-IDF), classification with the Naïve Bayes algorithm, model evaluation using a confusion matrix, and data visualization through word clouds. The results show that most user reviews are positive. The Naïve Bayes algorithm demonstrated good performance, achieving an accuracy of 80%, precision of 84%, and recall of 77%. Data visualization revealed frequently used words in positive reviews such as "excellent," "great," and "helpful," while negative reviews highlighted words like "please" and "driver." This study contributes to leveraging sentiment analysis to evaluate online transportation applications. Future research is recommended to explore alternative algorithms, such as K-Nearest Neighbor (KNN) or Support Vector Machine (SVM), and expand datasets with larger and more diverse data sources. These approaches are expected to enhance sentiment analysis model performance and provide deeper insights into user sentiment patterns..