Research On The Emotional Analysis Based On Set Pair-information Entropy In Social Network

Emotional analysis is the basis of hot spot mining,public opinion analysis and product recommendation.The Web2.0 era has formed a social network with rich emotional information on the basis of user participation,dominance and construction.Considering the texts in social network are short and mixed with emotions,traditional emotion-text analysis methods are less applicable.Based on the set analysis theory and the information entropy,the paper conducts an in-depth study on the emotion analysis in social network.The specific research contents are as follows.Firstly,the text is fully segmented to judge the combination confidence and the size of threshold to determine whether there is a new word discovery.The text segmentation is carried out by the Maximum Matching and Reverse Maximum Matching.The problem of ambiguity can be solved by the T-test of the tightness between word and word,which ensures the accuracy of the word segmentation to some extent.Secondly,the paper combines the Information Entropy,which is used to describe the degree of information disorder,and the Set Pair Theory together.It puts forwards the Set Pair-Information Entropy(SP-IE)algorithm to further analyze the diversity factors,and divides the text emotion of the individual into five categories: strong positive,weak positive,uncertain,weak negative and strong negative.The network platform provides users with a rich emoticons system.The emotional text and expression analysis of the users are carried out after translating the captured emoticons into the text meanings they carrying.The SPA-IE algorithm and HowNet algorithm are compared in terms of the accuracy rate,accuracy rate,recall rate and F value,which verifies the effectiveness of the analysis method.Finally,based on the SIR model of infectious diseases,an improved SIRE model is proposed by introducing foreign “immune users”.The affective analysis is predicted based on Python crawler and SIRE model.By comparing the fitting results and prediction results of SIR model and SIRE model,validate and analyze the effetciveness of the SIRE model.