Appraisal expression recognition based on syntactic path

Abstract

Different from previous works such as hand-crafted patterns and rule based methods, this paper proposes a novel method that uses syntactic paths to automatically recognize the appraisal expressions. In detail, the syntactic paths are first automatically collected to describe the relationships between the polarity words and their corresponding targets. Next, an edit distance based syntactic path matching strategy is used to improve the performance of the exact syntactic path matching method. Experimental results on the camera and MP3 player domains show that (1) syntactic paths are helpful for appraisal expression recognition, and (2) edit distance based syntactic path matching method can further improve the performance of appraisal expression recognition. © 2011 ISCAS.