Analysis of personality traits based on the disc model using machine learning methods

The analysis of a person's social media behavior with respect to privacy and human rights provides information about their personality traits and is seen as a topical task today. In areas such as marketing, training, education, human resource management and hiring policies in companies, knowledge about personality traits proves to be profitable and important in decision making and business orientation cases. The paper analyzes the performance of machine learning methods in a personality trait identification task based on the DISC psychological model and a small size dataset created from scratch. Although the dataset created was relatively small, the machine learning methods used showed encouraging and convincing results. Results for all personality trait classifiers were improved using hyperparameter optimization, increasing the performance of the XGBoost classifier to 70.45% on the accuracy metric in the test sets.