Detecting human sentiment from text using a proximity-based approach

Sentiment analysis seeks to characterize opinionated or evaluative aspects of natural language text thus helping people to discover valuable information from large amounts of unstructured data. Sentiment analysis can be used for grouping search engine results, analyzing news content, reviews for books, movie, sports, blogs, web forums, etc. Several methods have been proposed for sentiment analysis at word, sentence, and document levels mostly based on common machine learning techniques. In this paper we explore a new methodology for sentiment analysis by considering a new set of features based on word proximities in a written text. We propose three proximity-based features, namely, proximity distribution, mutual information between proximity types, and proximity patterns. We applied this approach to the analysis of movie reviews, drug reviews, and music reviews. Our experimental results show that proximity-based sentiment analysis is able to extract sentiments from a specific domain, with performance comparable to the state-of-the-art. To the best of our knowledge, this is the first attempt at focusing on only proximity-based features as the primary cues in sentiment analysis.