Research On Weibo Recommendation Method Based On User Interest

Micro-blog platform in social networking has attracted the majority of users' love and attention in recent years.It is understood that the platform will be added thousands of new users every day,and these users will leave thousands of messages on the platform at the same time.In the face of huge micro-blog information,the users are constantly searching for information which is consistent with their interests,so how to find the user's interest from the mass information and recommend micro-blog has become a hot issue in the present study.This article is based on this as a starting point,and studies problems that exist in the recommendation algorithm.First of all,aiming at the problem of low accuracy in user's interest mining,this paper proposes a micro-blog user's interest mining algorithm based on tag's update.Then,in order to solve the cold start problems in micro-blog recommended stage,this paper proposes a micro-blog recommendation algorithm combing tag and artificial bee colony.Finally,using the above two algorithms to design and implement a prototype system of micro-blog.The specific work of this paper is as follows:(1)Researched the algorithm of user's interest mining,and proposed a micro-blog user's interest mining algorithm based on tag's update.First of all,according to the various characteristics of the tag,to the user's initial interest is established.Then,calculate tag's update strength using the similarity and intimacy of the users with those which they concerned,and the influence of those which the users concerned.Finally,update the tag according to the tag's updating principle to establish the user's interest model.The method has a certain improvement in the accuracy and recall rate,indicating that this method is effective to express the user's interest.(2)Researched micro-blog recommendation algorithm,and proposed a micro-blog recommendation algorithm combing tag and artificial bee colony.First of all,the user's tag information is defined.Then,the fitness function of artificial bee colony algorithm is established by three variables,they are tag's weight,tag's preference and the similarity of the micro-blog words.Finally,artificial bee colony algorithm search strategy is used,the search has issued the optimal fitness value of micro-blog is found to recommend.This method can not only solve the cold start problem in the recommendation algorithm,but also has a certain effect on improving the accuracy of the recommendation algorithm.(3)Designed and implemented a prototype system of micro-blog recommendation based on user's interest.With the above two algorithms as the theoretical basis,this paper analyses and design the system modules and processes,and ultimately achieve a prototype system of micro-blog recommendation based on user's interest,users can discover and find your favorite micro-blog information.