Supervised Sentiment Analysis of Movie Reviews with SHAP-Based Interpretability Analysis
摘要
The use of technology allows users to post reviews of services and products on various online platforms. Users and service providers can benefit from a tool that helps them understand the sentiment behind the data they are presented with. The practice of detecting the emotional tone of the text contained in a document is called sentiment analysis. We implement several supervised sentiment analysis models that analyze IMDB movie reviews. Specifically, we evaluate machine learning and deep learning models such as Support Vector Machine (SVM), Logistic Regression, Naïve Bayes, Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), and BERT, RoBERTa, XLNet transformers. Further, we seek to enhance the performance of the best performing model by employing interpretability analysis using the SHapley Additive exPlanation (SHAP) framework, and contrast the results with other methods employed in sentiment analysis.