An Implementation and Analysis of Modified Approach for Mobile Apps Review Mining using Scrapper Package

Abstract

This study introduces a new method for mining mobile app reviews using the "scrapper" package and sophisticated ML algorithms. The goal was to improve app marketing and development by gleaning useful information from a massive database of user reviews. This was accomplished by classifying the retrieved reviews as either favorable, negative, or neutral using a thorough sentiment analysis. The outcome of this study laid the groundwork for future research on user preferences and perceptions. After that, we used a number of ML algorithms to refine the sentiment analysis and make the classifications more accurate. We used metrics like accuracy, recall, precision, F1-score, and confusion matrices are used for assessment of the performance of logistic regression, support vector machines (SVMs), neural networks, and Naive Bayes. Results showed that the suggested method worked, outperforming baseline methods in sentiment categorization by a wide margin. Another piece of evidence that the model could distinguish between good and bad feelings came from the ROC curve analysis. Finally, by presenting a strong and effective methodology, this research makes a significant contribution to the area of mobile app review mining. Researchers, app marketers, and developers all stand to benefit from a better understanding of user input and how to optimize app performance as a result of these discoveries.