Computing Personality Trait Based on Multi-source

Abstract

Personality computing is a hot research field recently, which mainly uses the individual’s traces on social platforms to understand, predict and analyze their behaviors, to make a certain accurate judgment of user personality types. The application of personality traits is of great significance to intelligent medical, personalized service customization, and other fields. The current personality computing research is mainly based on social media, the data source is limited, can not reflect the real situation. This paper collects college volunteers’ network access logs, the data is more comprehensive and complex, not limited to only the traces on social media. According to these network access logs, the calculation results show that the personality traits of users in different social media are not consistent, and the access preferences of network resources are not exactly the same as the personality traits reflected by social media. Therefore, it is more accurate to integrate multi-source network information to calculate an individual personality than a single data source.