Knowledge-Driven Approach to Predict Personality Traits by Leveraging Social Media Data

The day-to-day behavior of the individuals reveal their personality traits. With the emergence of the social media platforms, some aspects of this behavior are being recorded in their online profiles. This provides necessary input to develop algorithms that can predict personality traits of individuals. However, these algorithms need to exploit the semantics of the data in order to reveal the personality traits. Current studies on this topic mainly exploited the syntactic features of the language used by individuals to predict their personality traits. In this work we demonstrate the value of exploiting semantics of the messages conveyed in social media posts for predicting personality traits. In other words, we present a study that attempts to simulate the cognitive ability of the human brain, which allows to identify the important implicit information in social media posts for understanding the personality traits of an individual. Our approach shows the value of publicly available knowledge bases in eliciting implicit information in the user generated content and their impact on predicting the personality traits of an individual. We evaluated our approach using well-known 'myPersonality' dataset and showed that it outperforms the state-of-the-art algorithms that mainly depend on syntactic features.