The Research Of The Group Emotional Recognition And Evolution Based On The Customer Reviews

With the rapid development of e-commerce, online shopping has become the hottest trend in present China. While shopping online, consumers also can comment the user experience and product usage according to the network shopping process. Every comment is tinged with emotion, is a kind of emotional expression of the user, or the like, disgust, or opinions and Suggestions to the businessman, accurate and timely know these users mood is essential to the merchant’s business decisions. However, these huge number and unstructured information, user reviews focus on each individual mood is meaningless, also is not operating. Therefore, need to all user comments together, study mood change user groups, so as to effectiv ely support business management decision-making.This study is aimed at large B2 C e-commerce site user comments, made these comments information can effectively guide the business management decision, the concrete research content is as follows:(1) This paper reviews data preprocessing, and use the association rules to find the problem of frequent item sets from product features extraction. According to have been extracted to product characteristics, use the product key and emotional tendency word co-occurrence relation has realized "product characteristics- emotional tendency" word for extraction.Then, build the emotional tendency word corpus based on fuzzy number fuzzy theory, calculated the emotional tendency word fuzzy membership function and emotional polarity values. Using fuzzy cognitive map to "product characteristics- emotional tendency" word and its emotional polarity values for the knowledge representation, constructed the "product characteristics- emotional tendency" comprehensive analysis model.(2) On this basis, using the extraction to "product characteristics-emotional tendency" word on polarity and emotional value to build the source putted forward. Based on the evidence theory to the "product characteristics-emotional tendency" word to two fusion, has realized the group emotional recognition, and group emotional recognition experiment was carried out(3) Finally, the article is based on time sequence to achieve the evolution of the group emotional research, tracking the change process and trend of group emotional, this is the biggest innovations in this paper. And on the basis of group emotional evolution as a result, to guide the business management decision-makingThis paper introduce the concept of group emotional to user com ments mining, and based on time series research on evolution of group’s emotion, in the aspect of theory can further enriching the research in the field of network user comments; In practice, companies can make accurate and timely grasp the user needs, preferences and trends, the guidance of more efficient business management decisions.