Revealing Personality through Handwriting: A Fusion of Graphology and Machine Learning Techniques

This paper explores the integration of graphology and machine learning to analyze personality traits through handwriting. The research is motivated by the understanding that the brain expresses personality traits through neuromuscular movements, particularly in handwriting. By bridging historical graphological methods from the 19th century with contemporary machine learning techniques, this study utilizes a diverse dataset of 1108 handwriting image samples, sourced from Centre for Pattern Recognition and Machine Intelligence (CENPARMI) and a graphology expert. We employed machine learning algorithms such as k-nearest neighbor (k-NN), random forest, logistic regression, and transfer learning method, along with synthetic minority oversampling technique (SMOTE) for data balancing and ensemble methods like majority voting and stacking to classify and mine the images. Our experimental results indicate a significant improvement in prediction accuracy, exceeding 90% for traits like “Agreeableness” and “Open to Experience” using the stacking method. This research makes three key contributions: the innovative integration of graphology with machine learning for personality assessment, methodological advancements in handling imbalanced datasets, and the application of transfer learning in handwriting analysis. The findings illustrate the potential of this interdisciplinary approach to enhance personality trait prediction accuracy, offering valuable insights for psychology and personalized services. This study opens new avenues for future research in personality psychology and related fields.