Research On Consumer Online Shopping Decision-Making And Recommendation Of Commodity Based On Social Media Network

With the rapid development of the Internet and information technology, our country’s E-commerce market tends to the quickly development in recent years. Because of the virtuality of online transaction, the consumers’network trust is more important influence on consumer online shopping descion-making. Undoubtedly, the social media network application platform is the most convenient channels to promote the relationship and increase trust of the businessmen and consumer. Therefore, it is a great significance to explore what influence consumer online shopping decisions-making in the social media networks. Meanwhile, with the expansion of the network of information, it is difficult to find the useful information for the decision of consumer online shopping. And it is a challenging task for us how to effectively take advantage of the information to give commodities’recommendation with more satisfactory. So, recommendation technologies were presented on the need of solving the probelms, and the collaborative filtering recommendation algorithm is the most widely applied in many recommendation methods, but it also has some problems, such as the lower accuracy of recommendation. This dissertation is organized as follows:1. Consumer online shopping decision-making based on social media networkWith the development of the electronic commerce, consumer confidence is an important influencing factor of consumer online shopping decision-making. Based on the previous studies, this paper studied the cognitive ability, relationship and interaction strength of consumers in social media networks to affect on consumer online shopping decision-making from the network trust. And we made the quantitatively analysis of the key factors influencing consumer online shopping decision-making by using a multi-level regression method and sina-microblogging network’s real data.2. Recommendation method of commodity based on social media networkTraditional collaborative filtering recommendation algorithm efficiency has the problems of low cold-start and data-sparse, two new recommendation methods based on social network and social media were proposed in this paper. The first one analyzed the user trust in social networks and user preferences to get the nearest neighbor similarity sets and recommendate; the other one geted the trusted nearest neighbor set for targeted users to provide more accurate individual recommendation by integating similarities of users’ dynamic preferences based on users’ social tags, similarities between users based on users’ background information and the similarities of user rating based on time weight. Finally, we conducted experiment to verify the efficiency of the algorithm and the comparson to other recommendation algorithms by publicly available data sets.