Personality Prediction Based on Iris Position Classification Using Support Vector Machines

Abstract

Predicting personality generally involves personal interpretations of a person which makes the current methods for personality prediction process less adequate, timely and tedious. Thus, a simple yet efficient alternative method is proposed in this project for detecting iris positions which are used in Neuro-Linguistic Programming as clues for the human internal representational system and mental activity. This study set out to determine several positions of the iris of a person based on the Eye Accessing Cues. The design and the development of a complete system will be undertaken as for the users to use as a medium to predict their personality based on their iris position. Several pre-processing techniques were executed to each of the data before run into the testing and training activities for accuracy gaining. The algorithm used for classification of the positions is Support Vector Machine which by taking rectangle crop of an eye with 9000 pixels as inputs. Radial Basis Function is used for the kernel parameter of the proposed method. The classification will then map into the type of a person with the lists of his personality based on Visual, Auditory and Kinaesthetic theory. The result of the classification of the iris positions is currently 84.9% accurate which in the future might be increased by tuning several other parameters that consisted in Support Vector Machine.