Generating Sentiment Profile of Movies from Unstructured Textual Reviews

This paper presents an experimental framework for computing opinion polarity summary from free-form unstructured textual reviews of a movie. The system works on movie reviews obtained from the World Wide Web and computes a sentiment polarity summary in an automated manner. The experimental design is based on use of the SentiWordNet Lexicon. Two labeled movie review datasets are used to evaluate the design. The standard performance evaluation metrics of accuracy, f-measure and entropy are calculated to demonstrate the suitability of the proposed framework.