The Algorithm Reach Of E-commerce Personalized Recommendation Based On Web Mining Technology

With the rapid development of the network, B2C e-commerce develop veryquickly. However, e-commerce sites provide users with more choices at thesame time, inthe mass merchandise information, users often get lost and unable to find the product which they need. In the increasingly fierce competition environment of e-commerce，personalized recommendation systems are increasingly important,enterprises has become an important area of e-commerce research, it can simulate the sales staffrecommended product to their customers,so that customers can finish quickly shopping and enhance the competitivenessof the site according to their preferences for goods search.Based on the data mining technology, web mining, fuzzy clustering techniques,markov technology and e-commerce sites on the current personalized recommendation system, Researchers try to analysis and study personalized recom-mendation systembased on e-commerce web mining.First, the article of web mining technology research gives the main research contentsand methods at home and abroad, as well as personalized e-commerce web mining researchsignificance.Second, This chapter provides an overview of data mining technology, anddescribes the application of data mining and process.web mining techniques aredescribed in detail, including e-commerce web mining data sources, data chara-cteristics and web mining challenges.Third, the chapter analysis of the e-commerce site personalized recommendationtechnology,the traditional personalized recommendation technology and web-based miningfor personalized recommendation technology. The paper build a Web-based mining forpersonalized recommendation system, and analyzes the process of recommendation systemsfrom the offline and online recommendation process.Fourth, the chapter describes the main methods of fuzzy clustering analysis. The studyproposes clustering algorithm based on dynamic direct fuzzy clustering algorithm aboutusers clustering and pages clustering, and points out the advantages of the algorithm basedon the characteristics of web data. Fifth, this chapter provides the method which generates clusters by fuzzyclustering,then establishs corresponding markov prediction model to forecast among thedifferent types of clusters. The method greatly improves the accuracy of recommendation.Thereby effective measures for businesses can attract new users, maintain old customers andimprove enterprise competitiveness in e-commerce.Finally, according to the characteristics of web data the paper use fuzzy clusteringtechnique. The dynamic direct clustering fuzzy techniques is simple, fast, smallcomputational characteristics of the process of seeking to avoid the transfer matrix of fuzzysimilar closures. It applications in different category markov model to predict on the web forthe fuzzy clustering data mining, improves the prediction accuracy while reducing the cost ofcomputing time and space overhead. The paper summarizes the work inadequacies, as wellas prospects further research prospects.