Research on product feature extraction from Chinese customer reviews
Yongkui Zhang · Jisuanji gongcheng yu sheji · 2012
In view of the shortcomings of the existing methods which extracts product features from Chinese customer reviews,a template-based method of product feature extraction is proposed by using diverse techniques and resources such as lexical analysis,syntactic analysis and Tongyici Cilin to mine potential language knowledge in real corpora.Firstly this method conducts lexical analysis,syntactic analysis and artificial labelling to the reviews corpora,and then analyzes and sums up the global language rules from the marked results,and extracts the sequences of the POS and dependency parsing tags from the feature word and the opinion word,furthermore constructs the product feature templates by using Tongyici Cilin resource,finally the product features stated in consumer reviews are identified by using the feature templates.Experimental results show that the proposed method is effective.