Research And Realized Of Weibo Hot Topic Detection System

With the rapid and in-depth development of Internet,micro-blog and other social networks have been the main way for people to get the latest information.Finding hot topics in a large number of content in micro-blog is helpful to understand the message quickly.The traditional clustering algorithm for hot topics is not good when dealing with large,high dimensional and sparse data in the micro-blog platform,and the algorithm runs too long,and the result of clustering is far from the actual situation.In this paper,the problems found in the hot topic in micro-blog platform are studied.Aiming at the problem that K-means clustering algorithm has long consuming time and poor clustering effect in high dimensional and sparse nonlinear data set,the algorithm is improved from the direction of feature extraction.K-means clustering algorithm based on deep belief network.The algorithm uses the depth belief network to learn and adjust the data set continuously,to show the hidden structure and category attributes of the data,and then use the K-means algorithm to cluster the learning features to improve the effectiveness and efficiency of the clustering.Finally,the clustering algorithm based on deep belief network is applied to the hot topic discovery of sina micro-blog,and the hot topic discovery system simulation experiment is completed,which mainly includes three parts: data acquisition module,data preprocessing module and hot topic discovery module.The form of the key words is displayed.