Research And Implementation Of Personality Analysis In Cyberspace

Personality is a generalization of individual characteristics as well as a quantitative criterion to distinguish the difference between people.Personality is formed under the combined action of individual heredity,environment,learning and other factors,and has strong stability.Therefore,personality is often regarded as an important indicator of the research of"human"in various research fields.Personality is closely related not only to human emotion,language and behavior in real life,but also to human behavior in cyberspace.The ease of recording behavior in cyberspace and the rapid development of data mining technology make personality analysis in cyberspace feasible.And personality analysis gradually become a research hotspot.Using the data in cyberspace for personality analysis is of great significance to the personality analysis of traditional psychology.With the help of the personality analysis model constructed by machine learning technology,a large number of user personality information can be acquired in a short time,which can improve many services in cyberspace.Therefore,personality analysis in cyberspace has an important practical significance.In recent years,social network data has become an important data source for personality analysis.Social network status contains abundant user language information and has a strong correlation with user personality.Many studies use social network status to analyze users' personality.However,but most of them are conducted on inadequate label data due to the high cost of acquisition of label data.In this paper,to explore the usage of unlabeled data on personality analysis,a personality analysis framework based on semi-supervised learning is introduced.Besides,for making full use of the language information in social media status,the well-known n-gram model,LIWC and LDA is adopted to extract linguistic features.The experimental results demonstrate the semi-supervised learning ca:n take advantage of unlabeled data and improve the accuracy of prediction model.On the basis of personality analysis model,we builds a personality analysis system,including offline system and online system.Offline system completes the construction of personality analysis model and correlation analysis,and stores the model in the database for online personality analysis.Users can log on to the online system,analyze their personality,and view the results of correlation analysis between user language characteristics and personality characteristics.