The Research Of The Approach For Internet Commodities Comprehen-Sive Score Based On User Comments

User comments of commodity contain a wealth of information about its emotion,which is an important reference for shopping.With the coming era of web2.0,e-commerce platform emerges in an endless stream,the type and number of commodity and the users' comments grow rapidly,which is helpful for providing consumers with a variety of selections and abundant information of products.However,it is difficult for consumers to make a comprehensive comparison about quality,logistics,services,etc,which is hard for consumers to make the right purchase decision.Therefore,to help consumers make the right purchase decision,how to perform a comprehensive evaluation of Internet goods efficiently and accu-rately,is an urgent problem of the e-commerce currently.To solve this problem,this paper designed a method that using user reviews of the internet merchandise comprehensive scoring.The theoretical research and discussion in this thesis were carried out from the following three aspects:one is the analysis of existing reviews sentiment value calculation method disadvantage,this paper propose a menthod of comments scoring which based on the combination of N-Gram model,information gain method and logistic classifier;the other is presen-ting a menthod of internet commodity comprehensive scoring by analyzing the current review mining system shortcomings.This method extracted the product evaluation factors,calculation and evaluation of factors of weight vector and merchandise on each evaluation factor ratings to achieve a comprehensive rating Internet commodity.Third,through the experimentally verified reviews scoring method has higher accuracy and efficiency,The experimental results also show that:when the error coefficient is 0.1,the average of each evaluation factor goods accuracy rate of 90.31%.the design of This paper for the method of internet merchandise comprehensive scoring by using the user reviews of merchandise.This method can not only calculate a plurality of factors Internet commodity ratings and composite score which meeting the needs of individual users accurately that based on the pubic user reviews,but also provide supports of decision for the seller and consumers.Therefore,the application prospects of this paper are broad.