Detection of Personality Traits Through Handwriting Analysis Using Machine Learning Approach

The human emotions and behaviors differ from person to person. Human handwriting is a unique one like finger print, which carries the richest information to gain the insights into the physical, mental and emotional state of the writer. The science of analyzing handwriting is called graphology. Graphological rules are defined by considering the features of the handwriting. The proposed approach is carried out with the intention of finding how successful is the computer-aided handwriting analysis and the most appropriate machine learning classification approach for graphological feature classification. An application is developed to intake handwriting samples and to output personality trait of the writer. The application analyzes 3 major features: pen pressure, page margin and word size of a handwriting sample. Three hundred handwriting samples were tested, and the outputs were stored as a dataset. The dataset is used to train and test support vector machine and K-nearest neighbor models. The classification reports depict high accuracies of 96% and 85% for SVM and KNN models, respectively. Application was tested with a small team of candidates and received good responses. More than 65% of them were satisfied with the functionality and the output of the application. The proposed methodology is open for further enhancements one like handwriting sample can be segmented further into words and characters and can analyze the features of them.