Reserch And Implementation On Weibo User-Oriented Content And Friends Recommendation Algorithm

With the increase of network information,information overload becoming more and more serious.To solve the problems of information overload and low accuracy,we use social network,trust5singular value decomposition,text classification and so on.This paper introduce three recommendation algorithm and gives indepth study to three proposed algorithms.The main contents are as follows:(1)Research the recommendation algorithm based on social network,and propose a algorithm based on social network and trust. To solve the problem of low accuracy,this algorithm use social relationship and sequence similarity computing to dig the user’s similar community,and predict the user’s preference,at last recommend friends based on trust.(2)Because of there is no score data on the platform of Weibo,so we transform user’ action into score and propose a Weibo friends recommendation algorithm based on singular value decomposition.This algorithm uses the singular value decomposition to reduce the dimension of the user item rating matrix,combined with user similarity and project similarity to predict user’preferences.(3)Due to Weibo containing a lot of content information and user’ preferences are not unique,we propose a recommendation algorithm based on Weibo content and text classification.The algorithm mining the texts existing in the microblogging platform and classify the the user’s preferences documents,recommend different types of items to users,thereby improving the precison and recall rate of recommendation. At last,experimental tests show that three kinds of recommendation algorithm both in precision and recall rate with a certain degree of increase.