Ensemble Approach for Human Personality Classification using Textual Data

Abstract

Nowadays the users are active on social media platforms, such as Blogs, Twitter, Facebook, Instagram, etc. Users use these platforms to share their views on movie, current affairs, blogs, and group discussions. The views shared by users can be the charts, images, textual data, etc. The views or contents shared contain the inherent features of the users like, the persons they are following, the discussions on the topics. It will also help to provide insight of the personnel aspects of the users in the form of the job satisfaction, life expectations, the career preferences, psychological stage. Without the taken mandatory tedious tests or the feedback forms, the social media contents give the better classification of personality traits. Users share their views on the social media platforms as they are thinking about it without bothering about the tests taken by the organization. Personality classification process involves the preprocessing of the social media contents for feature extraction, mapping of the features to the personality model. Big five factor model is considered as the standards for classification of personality traits. In this paper we used the improved ensemble technique using gradient bagging for personality classification and result shows improved result as compared to recent methods of personality classification.