Yang Li, Wenjing Kong, Benchang Shia

【Abstract】

The study on the sentiment classification is challenged by the unbalanced, unmarked and non-standard web context data. In this paper, we proposes an adaptive semi-supervised topic-based classifierto figure the above issues.  Numerical study shows that the proposed method has strong adaptabilityto the imbalanced, unmarked datasets.  A sentiment classification of hotel comment context gains ef-fectiveness in predicting sentimental polarity of minority group in real study, which has confirmed the applicability and feasibility of this adaptive semi-supervised topic-based classifier in practical problems.

【Keywords】

sentiment classification, imbalanced data, semi-supervised learning, topic model