The Research Of Real Consumer Requirements Identification Of Online Personalized Product Design

Online sales and online personalized customization rapidly rise with the characteristics of convenience and low price in household and personal care industry, and are widely accepted by consumers, vendors, logistics and manufacturing industry. From the perspective of consumers, online history data, especially the ones with RUE, not only affects that consumers judge the performance of the products but also impacts on their purchase decision. From the point of manufacturing enterprises, they should make full use of online comments especially those with RUE, mine these data to identify the real-time and real consumers’requirements, improve manufacturing processes, product specifications, marketing and after-sales service model, form new and more personalized products for meeting the consumers’personalized requirements. Therefore, how to discriminate the reality of the comments, extract the features and how to judge the emotional tendency of the consumers have become the focus of enterprises and research hotspot in academia. However, the three problems still exist in the process of consumers’requirements identification:①More and more consumers without RUE can make comments that are with false user experience in Tmall and TaoBao.②In the process of choosing features, the method that choose features according to one word’s characteristic can’t merge the features whose meaning are similar or same, and reduce the number of features.③The research to determine the consumer’s emotional tendency in multiple negative comments is deficient. Taking these aspects into consideration, this paper puts forward the requirements extraction model based on RUE, homogeneity feature identification and mergence rules, and how to judge the consumers’emotional comments in multiple negative comments, and then takes the case of the comments in Tmall and TaoBao.1. RUE network comments extraction research. Write a program which is used to automatically acquire online comments.This paper combines Web text and Web log information, applies the ID which online comments publisher owns to obtain the corresponding log information and selects browse time, browse frequency and the number of comments from log as related parameters to present the rules that can judge whether the consumers have real user experience or not. And then eliminate comments information without RUE.2. Feature classification, conflation and extraction. Because different nouns or noun phrases are used to express the same meaning in the Chinese comments and the synonymous relation between Chinese characters are obvious, this paper presents: extract nouns or noun phrases as candidate features in the online comments, identify the homogeneity features by judging whether the candidate features conform to the conditions to be homogeneity feature or not, apply the merging rules to merge the homogeneity features, and establish the features mapping libraries to choose the features.3. The emotional tendency judgment in multiple negative comments. Negative comments contain the consumers’anticipation for product improvement and innovation, and it is also the important basis of manufacturing enterprise making product innovation. Aimed at the consumers often choose multiple negative words to express their feelings, this paper proposes three rules in multiple negative comments to judge consumers’emotional tendency for increasing the accuracy of identifying their requirements.4. Empiricalresearch. This paper takes a brand mobile phone in China as an example, applying the proposed methods to obtain the comments with RUE, extract the features and determine consumers’ emotional tendency, which realizes that how to distinguish the comments with RUE from the online reviews, feature extraction and how to determine consumers’emotional tendency. Compared the multiple figures, this paper verifies the methods and rules raised in the text are feasible and advantageous.