Aspect Extraction and Aspect Terms Expansion in Chinese Reviews Using Cluster Semi-Supervised Expansion Model

摘要

Aspect extraction is one of the most important tasks for text mining. Semi-supervised methods have been proposed to solve this problem. However, the seed terms have to be given in advance in these methods. The current methods categorize the aspects without expanding more aspects terms. And most of the methods are based on English corpus, there is a great space for the research on the aspect extraction method of Chinese Corpus. Therefore, in this paper, we propose a cluster semi-supervised expansion model to obtain aspect seed terms by clustering algorithm. The purpose of aspect seed terms is to find more aspect terms. Experimental results using Chinese reviews from three types of restaurants(coffee, hotpot, barbecue) show that the proposed method is indeed able to perform the task more effectively.