MBTI Based Personality Prediction o f a User Based o n Their Writing o n Social Media

Abstract

The undeniable power that various e-commerce and streaming websites exert on their users’ in terms of what they buy and what they watch is unquestionable, so the creation of better targeted advertisements and recommender systems is the need of the hour. Prediction of a person’s personality can be the key for the achievement of these goals. A novel way to understand the various facets of a person’s personality is by analyzing their MBTI (Myers–Briggs Type Indicator). This paper aims at classifying a user into any one of the sixteen personality types, defined by MBTI, through the use of natural language processing (NLP) and support vector machine (SVM) which was implemented on the MBTI dataset. Since the original dataset is unevenly distributed, SVM has been applied to the original dataset and an under sampled version of the MBTI dataset. The highest accuracy rate of 78.52% for the traits (thinking/feeling) was achieved in the original dataset whereas for the under sampled dataset it was 60.2% for the traits (judging/perceiving).