Thinking Like Psychologist: Learning to Predict Personality by Using Features from Portrait and Social Media.

Jie Nie,Lei Huang,Zhen Li,Chenxi Wei,Bowei Hong,Wenwu Zhu

DOI: https://doi.org/10.1109/ccis.2016.7790218

2016-01-01

Abstract:It is interesting but challenging to infer people's personality from multimedia data. This paper resolving this problem by jointly considered both human appearances from portrait image and behavior representative data from social media, where features are designed and selected separately inspired by psychology theory. Best features are selected for four personality categories respectively depending on classification performance. Experiment demonstrated that: by adding social media features, our method achieves an average improvement of 24.95% in true positive rate than the state of the art. Meanwhile, by introducing features inspired by psychology theory, our portrait features outperformed the baseline method by increasing 11.69% in true positive rate.