Related Technology Research On Micro Blog Sentiment Analysis

With the development of the network, microblog has a rapid progress. More and moreuser spread their information through microblog, the information of microblog has shownexplosive growth trends, how to acquire useful content from the massiceinformation,especiouly, how to analyze de hot topics and hot events is one of the mostvaluable problem. This paper carries out analysis and mining using micro blog sentimentanalysis technology. The researchment of micro blog sentiment analysis has good applicationand important practical meaning.A new feature selection method based on cloud model and combined with semantic wasproposed in this paper. This method improved the current2statistic feature selectionmethod combined with semantic through cloud model. By using the priority of cloud modelhandling uncertain knowledge, the distribution of features was amended. The experimentscarrird out on microblog data set verified the effectivety and feasibility of the proposedmethod.A novel sentiment analysis for microblog based on combinatorial Bayes network methodwas proposed in this paper. The different structure Bayes network classifiers can be seen as aseries of basic microblog sentiment classifiers and were iterated through boosting. In otherwords, every basic microblog sentiment classifier is trained on microblog training set. Thefirst basic classifier is trained using the original microblog training set and the other basicmicroblog sentiment classifiers was decided by the former basic microblog sentimentclassifiers. Those texts that were wrong decided will be appeared in the training set of newmicroblog sentiment classifiers. At last, all these microblog sentimen classifiers areconstructed to a combinatorial Bayes network. The experiments carrird out on microblog dataset verified the effectivety and feasibility of the proposed method.