Abstract：As a representative of the social network，microblog can help people find the realtime information that he is interested in easily from massive data by providing realtime and personalized information service.This paper proposes a personalized and realtime recommendation model for microblogs.In order to improve the efficiency and accuracy，PRT obtains users′ dynamic preferences based on LDA，divides users into different groups according to users′ preferences and provides the whole group of users with local recommendation lists.PRT recommends new microblogs to user，which is released from the time user browsed microblog last to the current time，and it meets users′ real-time and personalized need.The experimental results based on the real dataset from sina show the efficiency and accuracy of PRT.Compared with CT，the precision and recall of PRT increase by 10.6% and 8.4%，and it provides users with a more accurate recommendation.
石磊,陶永才,李俊艳,卫琳. 个性化微博实时推荐模型研究[J]. 小型微型计算机系统, 2016, 37(9): 1910-1914.
SHI Lei,TAO Yong-cai,LI Jun-yan,WEI Lin. Personalized and Real-time Recommendation Model for Microblogs. Journal of Chinese Computer Systems, 2016, 37(9): 1910-1914.
2016, Vol. 37 
