Improving sentiment analysis with Part-of-Speech weighting
Chris Nicholls, Fei Song · 2009
Sentiment analysis is concerned with classifying the opinions in a piece of text. We present a term weighting scheme which takes into account part-of-speech categories to improve machine learning-based classification of sentiment in product reviews. We experimentally find optimal strengths for each part-of-speech category and show that using this weighting method improves overall sentiment classification.