Sentiment classification of movie reviews using multiple perspectives

This study develops an automatic method for in-depth sentiment analysis of movie review documents using information extraction techniques and a machine learning approach. The analysis results provide sentiment orientations in multiple perspectives, each focusing on a specific aspect of the reviewed entity. Sentiment classification in multiple perspectives can provide more comprehensive sentiment analysis for applications like sentiment ranking and rating. By utilizing information extraction techniques such as entity extraction, co-referencing and pronoun resolution, the review texts are segmented into sections where each section discusses particular aspect of the reviewed entity. For each section of sentences, Support Vector Machine (SVM) using vectors of terms is applied to determine sentiment orientation toward the target aspect. In our exploratory study, we focus on the sentiment orientations toward overall movie, movie directors and casts in the movie. The experimental results prove the effectiveness of the proposed approach for sentiment classification of movie reviews.