Research And Application Of Micro-blog Friend Recommendation Algorithm Based On Big Data Analysis

With the rapid development of the Internet industry,all kinds of social network(SNS)platform has been rapid development,people's social activities from social networking to online social networking.Among them,the recommendation for users is an important online social networking services.In the social network,users can pay attention to the relationship,add their own interest in friends,expand their circle of communication,but how to recommend a high quality of concern for friends,has been one of the difficulties of personalized service.At the same time as the social network is increasingly hot,rising from massive data users caused by information overload,looking for offline friends,common enthusiasts and professional fields who bring great difficulty to the user,the recommendation of friends become the focus of attention of the user,and is also an important content of the research.The main contents of this paper are as follows:(1)In this paper,we propose a friend recommendation algorithm based on local random walk,and propose a recommendation algorithm based on trust degree and local random walk.The algorithm for walk in the choice of neighbor nodes is equal to the probability that the problem is proposed by combining user trust to predict the probability of selection of neighbor nodes algorithm,and according to this algorithm to establish a friend recommended candidate set.(2)This paper designs a friend recommendation algorithm based on time decay classification of interest,and improves the traditional recommendation algorithm.The traditional recommendation algorithm for the user to publish all the analysis of micro-blog,and then do the classification of interest,but did not take into account the user interest drift.In this paper,we introduce the time decay model for the problem of interest drift,and use the interest preference of the user's neighbors,modify the user's interest preference,and calculate the user similarity.(3)This paper designed a recommendation model based on classification of interest friends decay time,the model of candidate sets using the first algorithm,the user will be the candidate set published in the micro-blog data processing,feature extraction and classification of interest,finally combining the second algorithm to get friends to recommend a list of.The experimental results show that the proposed algorithm has a certain degree of improvement in accuracy and recall rate.(4)The real time recommendation system is designed,and the overall function of the system is completed.Through the design of the relevant database of the recommendation module,the real-time popular micro-blog recommendation module and the friend micro-blog recommendation module,the real-time recommendation system is implemented to help the user to recommend friends and select the real-time information.