Fatemeh Boustani - M.S. student, Dept. Computer and Informatics Payame Noor University Gheshm, Iran
Mehdi Khalili - Assistant Professor, Dept. Computer and Informatics Payame Noor University Tehran, Iran

With the rapid growth of artificial neural networks (ANN) structures in the diagnosis of different physical diseases, usage of these structures has increase to diagnose the mental disorders. In this paper, we propose an effective scheme to diagnose Schizophrenia using SVM and ANN structures: RBF, MLP and SLP. In the proposed scheme, the most suitable structure is chosen by applying the confusion matrix to all structures to achieve the minimum error rate in diagnosis. The experimental results for 150 hospital patient records with 24 features show that, the proposed scheme reaches to accuracy 98%, sensitivity 100%, specificity 97%, positive predictive value 96% and negative predictive value 100% in SLP structure.

Schizophrenia, Diagnosis, SLP, MLP, RBF, SVM