Personality Detection on Twitter Social Media Using IndoBERT Method

Abstract

Personality is the fundamental characteristic of human beings that makes humans unique. Because of these differences in human characteristics, personality becomes a benchmark for consideration in various recruitment processes. One way to predict personality is to apply an interview system or fill out questionnaires which often experience problems due to ineffectiveness in terms of time and cost. Results become inaccurate if prospective employees do not know themselves well. The big five personality method, divided into openness, conscientiousness, extraversion, agreeableness, and neuroticism, is widely used to predict personality. This study uses a deep learning method, IndoBERT, to detect personality based on five dimensions according to the big five personalities whose data is taken from Twitter tweets with crawling data. From the results of these studies, it is known that personality research using the IndoBERT method without a stemming process has a higher accuracy rate of 0.46.