Research on analyzing sentiment of texts based on semantic comprehension

Abstract

Text sentimental mining is aim to obtain the orientation information of text, and studying users' opinion is relevant because through it concerned departments could make correct responses. In this paper, we introduce methods to discriminatively learn semantic comprehension for use as features in text classification. Based on the existed HowNet Knowledge, we consider improving algorithm for calculating Chinese term semantic value. Then, conditions of adverbs, conjunctions and text-structure are added step by step. And the experimental results show that the proposed approach is suitable for judging texts' sentimental orientation.