Extraction of Chinese "Opinion target - Opinion word" Pairs Based on Part-of-speech Rules and Semantic Dependency Parsing

Due to the non-standardization of online comments, it is difficult to extract the "opinion target - opinion word" pairs. This paper proposes a method of extracting the "opinion target -opinion word" pairs based on part-of-speech rules and semantic dependency parsing (SDP). Firstly, adjectives, status words and some idioms are extracted as original opinion words according to the part-of-speech characteristics of opinion words, and then SDP is performed to obtain the final opinion words. Then, the opinion targets are extracted based on SDP and dependency parsing (DP), and filtered through part-of-speech rules. Finally, the implicit opinion targets are extracted based on probability. The experimental results show that the average F-Measure is 78.86%, which indicates that the extraction performance of this method is good.