Multi-classification analysis of large data based on knowledge element in micro-blog short text
Dong Yindi · International Journal of Computers and Applications · 2018
To solve the problem that the traditional emotion analysis method for microblog information adopting vocabulary target identification is unsatisfactory for identifying concrete concept application, a recommended system mixing behavioral habits of users and microblog user labels is proposed. Firstly, the content-based method is used to build a statistic model for emotion analysis of microblog information and experimental analysis was carried out for cluster coefficients of emotion analysis so as to acquire the best selected value of cluster coefficients; secondly, the decision classifier is used to build an emotion-based concrete method model for microblog information so as to realize the highest similar mode classification of microblog information and also use content and the method model to rectify it; finally, the performance advantage of the proposed method is verified in concrete statistic analysis through experiments.