Sentiment Analysis on Play Store Application Reviews Using BERT Model

Sentiment analysis is an essential task in natural language processing (NLP) that has various applications ranging from product reviews to social media monitoring. Traditional approaches frequently fail to capture the contextual nuances of language, which leads to poor performance. This research focuses on using a pre-trained BERT (Bidirectional Encoder Representations from Transformers) model for refined sentiment analysis of Play Store app reviews. The study explores tokenization, fine-tuning, and hyper-parameter optimization techniques to adapt the BERT model for sentiment analysis. Extensive trials were conducted to alleviate memory limitations by trimming the dataset and applying batching. The model's effectiveness was measured in terms of accuracy, recall, precision, and F1-score, resulting in accuracy of 84.3% and an F1-score of 0.84, outpacing traditional machine learning models and previous methods.