Human Personality Identification Based on Handwriting Analysis

Abstract

The study of handwritten patterns to recognise, assess, and comprehend a person's personality qualities is known as graphology. Handwriting analysis is pricy and error-prone, and its accuracy is dependent on the analyst's skill. The suggested method is therefore concentrated on creating a Web App with flask that can predict personality traits using machine learning without the involvement of a human. 657 authors' handwritten samples were used in the study as datasets. The user-uploaded image undergoes initial pre-processing, which removes undesirable properties, enhances quality, and applies transformations to make image data acceptable for feature extraction. We take into consideration seven manuscript patterns for filtering undesirable qualities, increasing the quality, and executing transformations for picture feature extraction. Eight Support Vector Machines are required to develop the unique personality and behaviour of the user after extracting all these features from the image. Predicting a writer's eight personality traits is fairly accurate. This includes emotional stability, mental acuity or willpower, humility, personal harmony and flexibility, lack of discipline, poor concentration, lack of communication, and social isolation classifiers with great accuracy.