The Hot Topic Detection And Correlation Analysis Based On Social Networks

In recent years, the rapid development of mobile Internet technology led to the second leap of the Internet. Rapid increase in the amount of data also accelerated the advent of the era of big data, people paid more attention to the value of data mining. The special social network application which is traditional and changing show the most of the charm of the data era. Since mobile communication technology developed to the 4G era, the social network is penetrated into every aspect of the lives of ordinary people. The value of commercial and research hidden behind the huge amount of data in the form of rich are waiting for people to explore, the results can be a huge impetus to the Internet industry forward.this paper through the analysis of the characteristics of the new era of social network data, summarize existing data analysis and processing method, puts forward a method based on social network hotspot information detection and correlation analysis. This paper first introduced the basic situation of the social network, analyzed the data characteristics of social network and data analysis of the focus and difficulty, summarized the basic process of social network data, and introduced the process and method of topic detection and correlation analysis.Specific research on technical methods were mainly divided into two aspects of segmentation and ranking. In the study of word segmentation method, based on the common summary and analysis of the Chinese word segmentation and new word recognition method, this paper improved the new word identification model according to the characteristics of social network data. The candidate strings are extracted through the Increasing N way, and then the new words can be screened out through the frequency, mutual information and information entropy of three aspects of filtration.This article used the word ICTCLAS tool combined with new improved word recognition model to segment word.In the study of ranking model, this paper summarized and analyzed the common ranking model based on user vote, combining with social network hotspot discovery request, combining the advantages of time slip model and the Newton cooling model, proposed a new mathematical model for the calculation of the rankings.The extracted hotspot Information were analysed on association rule by using Apriori model, researching the relationship between them, and the application of the results were partially explored.In the practice of model results, this paper had an empirical analysis of the above model based on Sina micro-blog platform, the results show that the research is feasible and efficient.Finally the research of the found of the social networking hotspots and correlation analysis in this paper are summarized,then this paper analysed the advantages and disadvantages of the methods and models, and the further research and development in the future was prospected.