Mobile Application Analysis and Forecasting App Ratings Before Usage with Hybrid Neural Network

The reviews and ratings play a critical role in app store platforms, enabling end users to gauge the performance and drawbacks of applications before installation. In this paper, we conduct a Comparative Model Evaluation using a dataset from the Google Play Store. Through the application of various models and algorithms, we compare their accuracy in predicting ratings. Subsequently, we implement a Hybrid Neural Network for feature extraction and leverage LightGBM for prediction due to its superior performance in terms of Mean Squared Error (MSE) compared to other models. Our study utilizes applications sourced from the Google Play Store dataset.