Research On Collaborative Recommendation Algorithm Based On User's Influence Measurement And Analysis

With the rapid development of HTML5,Web2.0 and other information technologies,and the research field has made many breakthrough results,which makes the application of Internet is becoming more and more popular,and its application range is more and more extensive.These developments,while giving more content to Internet users,also make the resources in the Internet grow at an exponential level,resulting in the problem of information overload.When users obtain massive information,they usually can not find the information quickly and accurately,resulting in waste of network information resources and the user's extra time consumption.In this context,personalized recommendation system came into being.At present,the personalized recom mendation system has become the most important method to solve the problem of information overload,and in the long-term application practice has achieved good results.In the personalized recommendation system,the collaborative recommendation algorithm i s the most widely used,because of its simplicity,high practicability and good performance.Based on this background,this paper proposes a collaborative recommendation algorithm based on user influence,and this algorithm is based on online social network,label system and cooperative recommendation algorithm.First,in order to improve the performance of the influence measurement algorithm,we propose a multi-dimensional influence analysis algorithm to predict the influence of users in online social networks.In the measurement module based on the user attribute,through the user's characteristic in the social network.In the influence analysis module based on the network topology,the central analysis algorithm is used to measure the network position of the user,and the network location of the user is measured by the centrality analysis algorithm.The comprehensive sorting module summarizes the measurement results of the previous two modules.The experimental results show that our proposed algorithm is s uperior to other influence measurement algorithms in terms of metric performance.Second,we propose a collaborative recommendation algorithm based on user influence.In contrast to existing research methods,our proposed algorithm integrates user preferen ces,user influence,and collaborative recommendation algorithms,and incorporates these into the proposed algorithm to achieve good recommendation performance.Experimental results show that integrating user influence into the collaborative recommendation algorithm,can lead a good result of microblogging prediction,users in the system can obtained better microblog resource recommendation.In addition,the collaborative recommendation algorithm based on user influence is also considered a better solution to the problem of data sparse.It is suggested that the collaborative recommendation algorithm based on user influence not only improves the recommendation performance,but also gives the user a better experience.