Research Of Personalized Recommendation Based On Online Shopping User Implicit Behavior

With the growing popularity of internet, e-commerce has obtained the rapiddevelopment. People staying at home can choose the goods or information that they ownneed, but inevitably it also brings users with the “information overload” problem.Personalized recommendation service is an effective way to solve this problem, and itplays a vital role in the sustainable development of the electronic commerce. The purposeof the article is to build personalized recommendation model based on user recessivebehavior characteristics, and put forward to improve the personalized service marketingstrategy, to improve the recommendation accuracy and efficiency, and enhancecross-selling capabilities, so that makes marketing strategy more targeted. According tothe characteristics of the subject, the research methods that this paper will use are themodel analysis method, comparative analysis method, empirical analysis etc.First of all, the research background and significance are detailed expounded. Thearticle compares and analyses current research situation of the domestic and foreignresearch on the data mining technology, based on the user’s personalized recommendationtechnology and the network marketing strategy, pointing out the existing deficiency ofpersonalized recommendation in e-commerce. The relevant basic knowledge of datamining technology,personalized recommendation and network marketing are carried onthe detailed elaboration, laid the theoretical foundation of this paper.Secondly, according to the existing problems in the e-commerce network marketing,puts forward the idea of personalized recommendation based on user behavior. In view ofthe actual situation that different influence factors have different contribution in theelectronic commerce, a multi-source matrix weighted association mining algorithm is putforward. combined with the algorithm, it constructed the personalized recommendationmodel based on user recessive behavior characteristics, and expounded the systemstructure of the personalized recommendation, effectively reduce the invalid mining,improve the quality of personalized recommendation service.Finally, based on user recessive behavior characteristics personalizedrecommendation model, analyzed and mining a instance, and combined with the example three points which are target user location, related product and associated web werefurther analyzed, put forward internet marketing strategy of improving the personalizedrecommendation service.