Sentiment Analysis of Google Play Store Reviews Using Bidirectional LSTM Networks

DOI

https://www.doi.org/10.59256/ijsreat.20260602023

Published In

Volume 6, Issue 2 (2026)

Published Online

March-April 2026

Pages

165-174

Article Type

Original Article

Check for updates

This study focuses on the development of an efficient sentiment analysis system for classifying user reviews from the Google Play Store using deep learning techniques. With the rapid growth of mobile applications, a massive volume of user- generated textual data is produced daily, making manual analysis impractical and inefficient. Traditional machine learning approaches such as Naive Bayes and Support Vector Machines often fail to capture contextual and sequential dependencies in text, leading to reduced classification accuracy, especially in the presence of complex linguistic patterns such as negation and mixed sentiments.To overcome these limitations, this research proposes a Bidirectional Long Short-Term Memory (BiLSTM) model, which is capable of processing textual data in both forward and backward directions, thereby capturing comprehensive contextual information. The dataset used in this study consists of 16,092 user reviews collected from the Google Play Store , where sentiment labels are derived from rating scores and categorized into positive, negative, and neutral classes. Prior to model training, the textual data undergoes a series of preprocessing steps, including lowercasing, tokenization, removal of stop words, and elimination of noise such as special characters and irrelevant tokens. The cleaned text is then transformed into numerical sequences using tokenization and padding techniques to make it suitable for deep learning models. The proposed BiLSTM model is trained using these processed inputs, and its performance is evaluated using standard metrics such as accuracy, precision, recall, and F1-score. Experimental results demonstrate that the model achieves an accuracy of 87.4%, showing a significant improvement over traditional machine learning models and standard LSTM architectures.The findings of this study highlight the effectiveness of BiLSTM in capturing contextual dependencies and handling complex textual patterns in user-generated content. This approach provides a scalable and robust solution for automated sentiment classification and can be further extended to real-time applications, recommendation systems, and large-scale opinion mining tasks.