Personality Analysis Based On Letter ‘t’ Using Back Propagation Neural Network

Abstract

This research is about personality analysis (graphology) based on offline handwriting using artificial neural network (ANN). Recognizing handwritten characters has been and still one of the most challenging problems in Artificial Intelligence (AI). Characters are rather complex patterns, having many variations in handwriting style. There are three objectives for this research. The objectives are to recognize small letter ‘t’ from set of handwriting and to identify the level of ambition of a person whether the person is optimistic, balanced or pessimistic. Questionnaire is used to collect handwriting samples and also been used to train the neural network. The questionnaires have been distributed to fifty respondents. The handwriting samples are taken from each respondent where each one of them writes one sentence of handwriting sample that used for personality analysis. Two neural network models are used in this research. The first model is used to recognize the letter ‘t’ and the second model used is to identify the level of ambition of the person. Both networks use back propagation algorithms for training. The first network model used two layer networks with 3600 input neurons, 75 hidden neurons and 1 output neuron. Meanwhile, the second network model used two layer networks with 3600 input neurons, 75 hidden neurons and 3 output neurons. Finally, the research shows that the result from both approach are equivalence and therefore, it proved ANN is a suitable tool for graphology. It determines back propagation neural network can be used to classify the letter ‘t’ and personality.