Text-based Explicit Personality Prediction And Personality Matching Analysis

Character is a person’s stable attitude towards reality,and the personality characteristics shown in the corresponding habituated behavior.A person’s personality has a huge impact on his or her life,affecting his life choices,preferences,and health.The automatic prediction of personality has many practical applications.In recent years,the automatic prediction of personality has attracted wide attention.There are two main problems in the field of automatic personality prediction:firstly,there is no suitable personality model in the field of automatic personality prediction;secondly,it is difficult to obtain data in the process of personality research.In order to solve the above two problems,on the basis of referring to various personality models and using basic psychological knowledge,a simple personality model----emotional personality model is put forward in this thesis.This model is a fourcategory model,which is based on the emotional factors of personality.Emotional factors are based on personality factors,which have universal applicability and can be evaluated by a third party without strong professional knowledge.In this way,the above two problems can be solved.In order to verify the model proposed in this thesis,We collected 247 subjects,and 300 microblog texts from each subject.Combined with the personality model of this thesis,a new text vectorization representation method is designed.Emotional values are obtained by scoring text emotions based on the sentiment dictionary,and emotional values are used as the vectorization representation of text.Based on the new text vectorization method,several classical machine learning algorithms such as decision tree to predict personality is used in this thesis.Compared with the classical text vectorization method Word2Vec,the results show that the prediction accuracy of convolutional neural network combined with the proposed method is the highest,reaching 92%,which also proves the rationality of the model.At the end,using the personality model proposed in this thesis,a personality matching algorithm is proposed for the application scenario of social networking.The influence of the communication process on the two people is simulated by the growth function model.By simulating two people to communicate,observing the changes in each other’s emotions before and after communication,then the personality matching degree is got.Finally,the data of eight people are used for verification,and the results are consistent with people’s experience.