Research And Implementation Of Public Opinion System Of Colleges And Universities Based On Sentiment Classification

With the rapid development of the network public opinion of colleges and universities has been further developed.More and more college students like to talk about personal ideals,express their personal opinions and vent their personal feelings on the Internet platform.Although the development of public opinion in colleges and universities has made college students have more right to express themselves,the rapid growth of the number of public opinion informations in colleges and universities has made management more difficult,which has caused a series of network crisis in colleges and universities,and has caused a serious negative impact.Based on the above problems,this paper constructs a set of public opinion system of colleges and universities based on sentiment classification.The system uses the web crawler to obtain the public opinion data of the colleges and universities.After the data preprocessing,the sentiment classification and related inquery are performed.For the research of sentiment classification,this paper uses sentiment dictionary and Naive Bayesian to judge the emotional tendency of public opinion texts of colleges and universities.This method uses the sentiment dictionary as a feature selection method,and constructs the public opinion texts of colleges and universities sentiment emotional classifier based on Naive Bayesian theory,then judges the emotional tendency of public opinion texts of colleges and universities.For the research of the related inquery problem,this paper uses the Latent Semantic Analysis algorithm to construct the word-document matrix for the public opinions texts of colleges and universities,and reduce the dimension of the matrix to construct the latent semantic space.In the latent senmantic space,the user's inquery statement is represented as a inquery vector,and the inquery vector is similarly calculated with other text vectors.Finally,the result is returned to the user according to the similarity value from large to small.The public opinion system of colleges and universities based on sentiment classification is of great significance for maintaining the security and stability of colleges and even society and promoting the healthy development of colleges and universities.The system has passed the system test and applied in the University of Jinan,and the running effect of the system achieved the expected.