Logistic Regression Model to Personality Type Prediction Based on the Myers–Briggs Type Indicator

Abstract

The Myers-Briggs Type Indicator (MBTI) is a widely-used tool in psychology for determining personality types, playing a crucial role in fields like team building, communication, and personalized marketing. Despite its popularity, accurately classifying MBTI types using machine learning remains a significant challenge. This study focuses on addressing this challenge by exploring the effectiveness of logistic regression in MBTI classification tasks. Two approaches are used: four-times binary classification and multi-class classification. The findings show that while logistic regression performs exceptionally well in binary classification tasks but the accuracy is not good in multi-class classification. Additionally, combining binary classification results yields an overall accuracy that is lower than the direct multi-class classification. These results highlight the limitations of logistic regression in multi-class tasks and suggest the necessity for more advanced models. Future research should focus on improving multi-class classification accuracy, potentially through more complex architectures or hybrid models combining binary and multi-class approaches.