Research On Recommendation Technology Based On Comprehensive Tag Factor And Clustering Algorithm

With the rapid development and popularization of Web2.0 technology,the Internet data shows an explosive growth.Recommended system can help users efficiently find the resources that they are really interested in.The recommendation system solves the problem of information overload,but also faces the problem of data sparsity and cold start.The key to the success of the recommended system is the understanding of the user's interest,usually the user's interest is difficult to describe the contents of the user itself,the tag system is through the use of rich feature information to find the user's interest in the search for resources.Therefore,based on the research of the traditional labeling system,the recommendation system and the clustering algorithm,this paper takes into account the various factors of the tag and combines the clustering algorithm to carry on the traditional tag-based personalized recommendation algorithm The optimization,so that it can flexibly cope with sparse data,cold start and timeliness and other issues,and thus achieve the purpose of improving the recommended quality.In order to solve the problem of data sparsity and cold start in the traditional recommendation algorithm,the user characteristics are introduced on the basis of the traditional tag-based personalized recommendation algorithm,and the users are clustered by using the user characteristics,alleviate the cold start problems.In addition,the user interest will change over time,so the algorithm adds a time attenuation factor,according to the tag of the different labeling time to give different weights.At the same time,in order to quickly find the same interest with the target user and easy to be accepted by the user,enhance user identity,improve recommended efficiency,the algorithm introduced a trust relationship and user preferences.The experimental results show that the proposed algorithm can effectively improve the recommendation accuracy.Finally,aiming at the poor scalability problems of traditional recommendation algorithm in dealing with massive data,parallel design algorithm based on Hadoop platform,improved the computational efficiency and has good scalability.