Improvement of a Naive Bayes sentiment classifier using MRS-based features

Abstract

This study explores the potential of using deep semantic features to improve binary sentiment classification of paragraphlength movie reviews from the IMBD website. Using a Naive Bayes classifier as a baseline, we show that features extracted from Minimal Recursion Semantics representations in conjunction with back-off replacement of sentiment terms is effective in obtaining moderate increases in accuracy over the baseline's n-gram features. Although our results are mixed, our most successful feature combination achieves an accuracy of 89.09%, which represents an increase of 0.76% over the baseline performance and a 6.48% reduction in error.