Key points are not available for this paper at this time.
Online social networks are becoming a very rich source of user generated content. This content motivates different types of applications that rely on personalization; such as recommender systems and online marketing. Detecting personalities through mining publicly available social data immerges as an important related issue that can assist web-based systems. Some approaches have been introduced to use publicly available social data to infer user's personality. This paper presents an approach for personality traits inference based on text semantic analysis. Different representations of user text combined with several semantic based measures are proposed to predict users' personality through their Facebook status updates. The proposed approach has been tested and validated on data released by the myPersonality project for the Workshop on Computational Personality Recognition. The results prove that the information content-based measure achieves the best average personality trait prediction with an accuracy of 64%.
Hassanein et al. (Sat,) studied this question.