Optimization of Sentiment Analysis for Amikom One Application Reviews Using SMOTE with Artificial Neural Network Algorithm

Sentiment analysis is a technique to decipher textual views and feelings. This study assesses a model's performance in sentiment analysis of Google Play Store reviews for the Amikom One app. With more unfavorable ratings, the primary problem is the imbalance in classes. It was done using the Synthetic Minority Over-sampling Technique (SMOTE) to remedy this. The techniques used are preprocessing the data, using SMOTE, and classifying sentiment using an artificial neural network (ANN). F1-score, recall, accuracy, and precision are used in the model evaluation process. The outcomes demonstrate a great degree of accuracy improvement in the ANN model's performance following the use of SMOTE. On training data, the model successfully classified sentiment reviews with 100% accuracy, while on test data, it achieved 93.44% accuracy. Sentiment research shows that 54.10 percent of the evaluations are favorable to the application, with 45.90% being critical. This study Artificial Neural Networks' (ANN) potential in sentiment analysis of mobile application reviews, offering developers with useful insights into how to enhance program quality using user feedback.