Research On Weibo Recommendation Algorithm Based On Tag Clustering And User Preference

With the rapid development of social networks,microblog has been favored by many users with its characteristics of “short,spiritual and fast”.As the number of microblog users continues to increase,microblog data has also exploded.Faced with so many complicated microblog data,users often spend a lot of time and energy to find the microblog information they are interested in,and the user's preferences are diverse.Therefore,how to analyze user preferences from these data and recommend the microblogs of interest to users is one of the hot issues in social network research.In this paper,the problem of low quality of user tag generation is considered.Considering the user tag feature and user social relationship,a microblog recommendation algorithm based on tag clustering and user preference is proposed.A microblog recommendation prototype system is designed and implemented.The specific work is as follows.1?A microblog user tag generation algorithm based on TextRank and Word2 vec is proposed.This paper studies the relationship between user tags and user's microblog content,analyzes the insufficiency of user tags,considers the word part of speech,adjusts the initial weight of words,and then combines TextRank algorithm and Word2 vec model to generate user tags for microblog content.2?A microblog recommendation algorithm based on tag clustering and user preference is proposed.Considering the characteristics of user tags and user social relationships,firstly,the density peak clustering algorithm is used to calculate the user preferences of the microblog user tags,and the PageRank algorithm is used to obtain the attention of important users to indirectly obtain the similarity of important users,and then The two are weighted to obtain the user's comprehensive preference similarity,and the microblogs that the user is interested in are recommended to the target user.3?Design and implement a microblog recommendation prototype system based on tag clustering and user preferencesBased on the label clustering and user preferences,the design principle and function modules of the microblog recommendation system are analyzed.The microblog recommendation prototype system based on label clustering and user preference is designed and implemented with good results.