Personality Prediction of Social Network Users using LSTM based Sentiment Analysis

The rapid increase in posting personal information on social media has been noticed over last two years especially due to heavy use of smart phones and restrictions imposed on face-to-face meeting due to the current Covid-19 pandemic. Often, user information and user status updates on social media platforms provide valuable insights in determining personality traits. Scraping and analyzing this social media content and accurate prediction of user personality bring many benefits for enterprises including delivering personalized services, product recommendation and improving performance of recruitment system. This study was carried out based on digital footprints left on social media and the Myers-Briggs Type Indicator (MBTI), one of the most reliable personality inventories, was used to identify and predict the personality traits in mobile applications. Long Short-Term memory (LSTM) architecture-based sentiment classifier was proposed in this paper to predict personality and obtained a convincing model performance with overall 78% accuracy for personality traits.