Research On The Evaluation Of Business Reputation Based On UGC In Social Commerce

With the rapid development of the Internet,the social media,such as We Chat,microblogging,social networking sites,has also been rapid development.At the same time,resulting in the integration of social media e-commerce,that is social commerce.In the social commerce environment,the consumer as the main producer of the text data was explosive growth,the major social commerce platform,there have been a lot of text reviews,merchandise recommendation,commodity purchase and sharing information,these information collectively referred to as user generated content,that is UGC,mainly refers to the user's original content,including text,pictures,video and other content.And the user-generated content contains a variety of factors related to the evaluation of the reputation of the business,so the user-generated content of the text data extraction,mining,and out of business reputation evaluation related to the key information,which is conducive to promoting social commerce and provide strong support for consumer purchasing decisions.This paper mainly through the text mining technology,the user-generated content of the keyword extraction,keyword clustering and cluster-like tag extraction,and then generate business reputation dimension and build credibility of the credibility of the index system,and finally evaluate the business reputation level.The main contents are as follows:Business reputation dimension extraction.This paper proposes a short text subject extraction method based on HNC and lexical chain.In the process of constructing lexical chain,combine the word similarity based on HNC and the word relevance based on word co-occurrence model,propose the calculation method of semantic fusion degree between words,and measure the semantic relation between candidate subject words and lexical chain by semantic fusion value.After the construction of the lexical chain is completed,the comprehensive weight of all the lexical chains is calculated,and the words in the lexical chain with the combined weight greater than the specified threshold are selected as the keywords.Through the keywords cluster,cluster labels extraction and cluster labels are summarized artificially,we can get the corresponding business reputation dimension.Finally,text data of Mogujie e-commerce platform are used for experimental and case analysis,which proves the feasibility and superiority of the proposed method.Business reputation evaluation.Constructing the reputation evaluation index system based on the generated reputation dimension,by introducing the cloud model,this paper proposes a comprehensive evaluation method of business reputation based on cloud model.In the evaluation process,the cloud model is used to construct the comprehensive evaluation matrix of credit index,and combined with the weight of the reputation index,the business reputation comprehensive evaluation cloud model is obtained.The similarity between the comprehensive evaluation of the cloud model and the set of cloud models is calculated,similarity of the largest collection of cloud model corresponding to the credibility level is the business reputation level,and the evaluation results in the form of cloud drops intuitive display.In this paper,business reputation dimension extraction method and business reputation evaluation method are proposed,promoting the further development of relevant scientific research,the existing problems of the business reputation evaluation in the social commerce have been improved,it has a certain practical significance and practical significance.