The paper presents methods of retrieving blog posts containing opinions about an entity expressed in the query. The methods use a lexicon of subjective words and phrases compiled ...
We present a general learning-based approach for phrase-level sentiment analysis that adopts an ordinal sentiment scale and is explicitly compositional in nature. Thus, we can mod...
This paper presents Q-WordNet, a lexical resource consisting of WordNet senses automatically annotated by positive and negative polarity. Polarity classification amounts to decide...
Sentiment classification refers to the task of automatically identifying whether a given piece of text expresses positive or negative opinion towards a subject at hand. The prolif...
This paper demonstrates that the imbalanced data sets have a negative effect on the performance of LDA theoretically. This theoretical analysis is confirmed by the experimental r...