An Improved Method to Building a Score Lexicon for Chinese Sentiment Analysis

This paper presents an improved method to automatically generate and score a new sentiment lexicon for Chinese sentiment analysis. Node represents word and weighed edge stands for the relationship between two words in a weight graph. The polarity detection is achieved through the label propagation in this weight graph. Moreover, the lexicon is expanded through some direct relations between the words in a thesaurus. The experimental results demonstrate that the proposed algorithm can not only effectively improve the accuracy of the approach for constructing a sentiment lexicon, but also obtain relatively accurate scores. Meanwhile, the lexicon also guarantees the wide domain coverage.