STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Communication Engineering & Systems

Personality and Behavior Identification Based on Handwriting Analysis

  1. Maahi Khemchandani
  2. Niharika Shailesh Ghali
  3. Disha Dinesh Haldankar
  4. Rahul Kiran Sonkar

Abstract

Graphing is the process of identifying, evaluating, and understanding a person's personality traits through handwritten patterns. The accuracy of handwriting analysis depends on the skill of the analyst, it is expensive and prone to errors. The proposed approach is therefore focused on building a system that can predict personality traits with the help of machine learning without human intervention. In this project, 657 authors' handwritten samples were taken as datasets. The image uploaded from the user is first pre-processed which makes image data suitable for feature extraction by filtering unwanted attributes, enhancing the quality, and performing transformations. We consider seven handwriting patterns, for image feature extraction by filtering unwanted attributes, enhancing the quality and performing transformations. Image features used for extraction are: (i) Baseline angle, (ii) Top margin, (iii) Letter size, (iv) Line spacing, (v) Word spacing, (vi) Pen pressure and (vii) Slant angle. After extracting all these elements from the image, eight Support Vector Machines are trained that produce the individual personality and behavior of the user. Emotional stability, mental energy or will-power, modesty, personal harmony and flexibility, lack of discipline, poor concentration, non-communicativeness, and social isolation are the eight personality characteristics of a writer that can be predicted with high accuracy.

Keywords

Support