Research On Establishing Semantic Dictionary Based On Book Reviews Analysis

Books play an important role in everyone's learning career, there are a large number of online book reviews, how to obtain key information from the complex comments become the focus of attention of scholars and experts. The machine automatically processed natural language understanding needs the support of dictionary. Existing HowNet, WordNet and other authorities semantic dictionaries are all common dictionaries.Maybe its profession is not strong and it's less efficient. This paper is to solve these cases, combined dictionary structure principle of HowNet and WordNet, study and propose to build a method-specific semantic dictionary based on book review analysis, it can be used in the subsequent book review work.The main contents of this article are completed:According to the most four common websites containing book information, analyze their web and data features, choose Dangdang and Jingdong as the data source of book review, improve existing crawlers and choose the correct strategy to get a large number of book reviews, store them in the database with format.According to the thought of semantic dictionary construction, complete data cleansing, retained valuable comments data, complete Chinese word segmentation and POS tagging, set the extraction rules, extracting the desired high-frequency words.According to the speech and semantic of the vocabulary to finish classification of high-frequency words, put them into the corresponding classification dictionary, build the expansion of vocabulary and semantic structure dictionary based on the relationship between the words, use the semantic dictionary to analyze existing text books reviews to verify the validity of semantic dictionary.In this paper, collected more than 30 million of the book reviews and finish information research and data processing, proposed the method of special book semantic dictionary construction, after experimental verification, the dictionary can be used to obtain simple book review corpus analysis, provides an effective data support for the subsequent evaluation of Books.