Fake App Detection Using Sentiment Analysis

In today's environment, fraudulent actions in mobile applications have become a major worry. It is now critical to identify and stop fraudulent activity within mobile apps due to their continued rapid growth across multiple sectors. This abstract presents a comprehensive approach to fraud app detection leveraging sentiment analysis techniques. Our goal is to identify fraudulent mobile applications within app stores by analyzing user-generated reviews and ratings. We first compile a dataset of app reviews and get the text data ready for analysis. Sentiment analysis is employed to assign sentiment scores to each review, providing insights into user opinions. We present a machine learning model trained on labeled data to identify whether an app is fraudulent based on the sentiment scores and other relevant criteria. Standard measures like F1-score, ROC-AVC, accuracy, precision, and recall are accustomed to evaluate performance. A visual evaluation of Visualizations such as ROC curves and precision-recall curves demonstrate the model's ability to discriminate between fraudulent and non-fraudulent apps. Our results shed light on the effectiveness of our fraud detection system, enabling informed decision-making and potential improvements. This project emphasizes the importance of data-driven approaches in addressing the ongoing challenge of app fraud detection.