Research On Emotional Tendency Based On Microblog

As an important network social platform,micro-blog is deeply loved by people because of its high real-time and convenient sharing of information.In recent years,more and more people express their feelings or opinions through micro-blog.Many hot events also come from micro-blog,so micro-blog has become the focus of Internet public opinion,and has also become a research hotspot in recent years.The network public opinions deeply affect people's lives,but also affect the social stability and security.The analysis and application of the network public opinion through technical means is of great significance to prevent the occurrence of major events and make the government make a quick decision.This paper analyzes the background and significance of the research on the emotion tendency,and analyzes the current research status through reviewing the literature at home and abroad,and summarizes the opportunities and challenges of the current research.This paper analyzes the text features of micro-blog,sorts out the commonly used emotional dictionaries,and extends the emotional lexicon combined with micro-blog data,and constructs a more comprehensive emotional dictionary.This paper also introduces the commonly used text categorization algorithm,and discusses the use scenarios and advantages and disadvantages of the algorithm.By improving the feature selection method and analyzing the semantic features,we get good results in this paper.Based on the background of today's data era and the characteristics of micro-blog,this paper carries out experiments from different angles,and uses two different methods to determine the sentiment tendency of micro-blog.From the view of machine learning,aiming at the lack of sentiment analysis method of feature selection process,this paper proposes a multi feature fusion factor selection methods,and the integration of the existing emotional dictionary,self-learning from micro-blog in the text through the emotional word recognition algorithm,to construct a more comprehensive for micro-blog the text sentiment dictionary.Through the intervention of the emotional dictionary in the process of feature selection,the quality of the selected features is better.The weight of feature is allocated by feature weight algorithm,and the data that conform to the classification format is constructed.Finally,the sentiment tendency of micro-blog text is determined by classification algorithm.From the semantic point of view,this paper analyzes the semantic composition of chinese text in detail,and analyzes the composition of the sentences in the text and the lexical character of the words.By constructing polar dictionaries and self-defining emotion computing rules,we set up the calculation formula of text sentiment orientation judgment,and use the calculated emotional values to achieve the sentiment classification of texts.The calculation method proposed in this paper makes up for the shortcomings of the existing methods,enriches the rule of emotion judgment,and realizes the accurate judgement of the sentiment tendency of micro-blog text.Through experimental verification and result analysis,the feature selection method and affective computing method proposed in this paper have achieved good results in the set of comparative experiments.In the emotion lexicon fusion machine learning method,two sets of contrast experiments are set up respectively: word frequency method and mutual information method.Through data visualization technology,we can clearly see the superiority of the feature selection method in this paper.In the affective computing method,three sets of comparative experiments are set up.The experimental results show that the polarity dictionary built in this paper is more comprehensive,and the proposed affective computing method has better experimental results.