Predicting Personality with Twitter Data and Machine Learning Models

This study applies personality prediction of a Twitter user based on the words used in tweets posted by the user. The personality type is predicted based on Big Five Personality Model that outputs agreeableness, conscientiousness, openness, neuroticism, and extraversion as personality traits. We analyzed Turkish words for prediction, prepared a new dictionary that includes Turkish words with their special word groups. The most successful machine learning methods are selected to predict each personality trait. When the machine learning models were trained with the latest 50 tweets of users, models estimated each personality trait with the accuracy values in the range of 0.76 to 0.97