Identification of Big Five Personality on Twitter Users using the AdaBoost Method

Abstract

Social media is one of the lifestyles in the modern era that uses web-based technology for social interaction. As one of the most popular social media platforms, Twitter allows users to express themselves through tweets that can show their personality. The Big Five theory states that a person’s personality is divided into five dimensions: openness, conscientiousness, extraversion, agreeableness, and neuroticism. Several methods have been used to conduct user personality research based on activity on social media. The AdaBoost method is used in this study to identify the personality of Twitter users using sentiment, emotion, social, PCA, and POS-tag features. There are two test scenarios in this study. The first is testing the AdaBoost model with all features, and the second is testing the AdaBoost model with a combination of three features. The research indicates that the data preprocessing method can affect the model. The results showed that the AdaBoost model with all the features and without the stemming process had the highest accuracy value of 53.57%.