Automatic Identification of Personality using Writing Behavior

Abstract

Handwriting is one of the most common and oldest expressions of human behaviors. We learn to write by copying shapes from a standard set of pro- totype alphabets the copy book. With time, every individual deviates from the scholarly writing style as a function of personal preferences. These prefer- ences include shapes of drawing the letters (allographs), presence or absence of connection between two letters, word and line spacing etc. This allows handwriting to be used as a biometric modality to identify the author of a questioned document or verify the authenticity of a document or a signa- ture [11]. Moreover, studies have shown that handwriting can be e ectively employed to predict a number of demographic attributes of writers including age group, gender, race and handedness etc. [12]. Handwriting can also be used as a non-invasive diagnostic tool for prediction of a number of neuro- logical disorders including Parkinson [10] and Alzheimer [13]. In addition, psychologists and graphologists exploit handwriting and hand-drawn shapes to identify the personal attributes of an individual [1, 2].