Business Reputation Analysis And Evaluation In Social Commerce

Social media and electronic commerce is infiltrating the other fused with each other widely, which produces social commerce possessing dual attributes. Under the social commerce, the rapid growth in the number of business and the uneven level of business reputation make research on business reputation evaluation more urgent and important. Comment texts are the treasure contained in the merchant’s reputation information. Through text mining technology like extraction and semantic analysis, key information hidden can be quickly obtained, helping consumers make purchase decisions and helping businesses improve service level and credit level. Aimed at above issues, in this paper, keywords are extracted, and clustering analysis and sentiment analysis method are adopted to build the business reputation dimension system and evaluate the level of business reputation. The main contents of this paper are as follows:Building the business reputation dimension system. Keywords extraction method for comment texts is proposed, which is divided into three steps including preliminary extraction, keywords expansion and frequency adjustment to extract completely. Keywords clustering algorithm based on HNC is put forward, introducing HNC to CURE algorithm to improve clustering effect. Then representative points are used to describe clustering results as the name of reputation dimension, keywords frequency is utilized to calculate the weight of each dimension. Finally, experimental results show that the accuracy of proposed keywords extraction method and keywords clustering method. Taking Jingdong platform mobile phone review text as an example, the business reputation dimension system is built up, which proves the feasibility and validity of the method.Business reputation evaluation. Firstly, restoring the keywords in the reputation system to the comment texts, the emotion words, degree adverbs and the negative words modifying keywords are located by the dependency syntax. Secondly, emotional value calculation method of emotion words based on HNC, formula for calculating the weight of negative words based on distance and degree adverb weight table are put forward to measure the emotional value of keywords. Thirdly, keyword frequency is applied to determine the reputation value of each dimension, and reputation dimension weight is used to calculate the value of the business comprehensive reputation. Finally, experimental results prove the congruency between the proposed method and the manual annotation in emotional value calculation. In case analysis, we analyze and evaluate the reputation level of the business through calculating the reputation value of the business reputation dimension system in the case of last section.The business reputation dimension system construction method and the reputation evaluation method enrich the research of text mining theory, make up for the lack of business reputation evaluation under the social commerce background, and have good application prospects.