A Study On Chinese Comment Opinion Mining Approachbasedon Catering Domain

Catering industry as the market size of traditional mainstream industry also stood in the change of \ "discourse\".According to the relevant departments of statistics,in 2014,domestic supply of raw materials field for more than eight hundred billion yuan,and food consumption area reached three trillion yuan,the catering industry chain upstream and downstream from personnel of course of more than 100 million people,the catering industry has become China's real estate,catering industry's largest industry.Catering industry as the industry has enormous influence on the word of mouth,the tendency of customer praise and colour can directly affect the actual income and other customer choice,so for food industry evaluation of the emotional attitude,both to users and to businesses,its meaning is very profound,user evaluation by means of analysis of related products,to better understand the product,make itself determine when do you want to buy more scientific;In addition,businesses can use practical evaluation of user study,constantly improve product quality,to establish reasonable management strategy,make the company's economic profit rose sharply.Evaluation on the Internet,however,the number is large,and every day will appear many new comments.Although numerous and complicated information can provide us with great convenience;However sometimes vast amounts of information,also bring us a lot of trouble.For example,a user want to understand a product usage or vendors want to first find out the market marketing present situation and improve the quality of products,such information search,screening,rejecting,consolidation,this job is very complicated and waste time.In this paper,the text reviews the new field is introduced into the food and beverage,emotional text use of catering industry,established a set of suitable for food industry knowledge base,found out the most suitable for food industry comment text theme identification technology,and through the comparison with machine learning,advancing the Chinese comments mining accuracy of syntactic analysis method.In this paper,starting from the second chapter mainly introduces the opinion mining research review,the concept of opinion mining,network new media concept,intrinsic meaning of the text orientation analysis and the analysis of the concrete task content,covering the subjective and objective classification,views holders,evaluation subject,emotional words extraction,and tendentiousness recognition,sentence orientation analysis,etc.,at the same time also introduces the basic concept of opinion mining,corpus construction;The third chapter,mainly about emotional dictionary building restaurant reviews,including common emotional dictionary,weibo language dictionary sets,NTUSD dictionary,emotional expression symbol dictionary,the phrase dictionary,network hot word dictionary,catering to evaluate emotional dictionary,etc;In the fourth chapter is mainly based on the CRF model evaluation subject dining areas,by CRF model,to find the effective feature of the model,and then,according to the data of the sample,do the actual food evaluation to identify the theme of the experiment and verify the accuracy and feasibility of CRF model;Finally,mainly for catering comments orientation analysis,first of all food and beverage comments for sample analysis,according to the results of sample analysis,as well as the analysis of characteristics of model to build the restaurant reviews emotional vocabulary,according to the emotional vocabulary,complete meal evaluation subject orientation analysis,and through the Chinese commentary orientation study of food service industry,to obtain a tendency for catering services segment and effective classification method,and builds a catering evaluation ontology knowledge base,with a comparative evaluation of emotion tendentiousness algorithm of complete food and beverage service.In this paper,in the sixth chapter will be in front of the experimental model of modular encapsulation,implement systematic calculation and display,through the feasibility study,we hope can continue optimization and improvement of gradual application value and achieve bigger and bigger.