Abdollah Ansari - M.S. Student, Department of Computer and Informatics, Payame Noor University, Tehran, Iran
Mehdi Khalili - Assistant Professor, Dept. of Computer and Informatics, Payame Noor University, Tehran, Iran

Factors such as the increasing prevalence of diseases, the need for ongoing monitoring of disease progression, lack of access to specialists in some areas, saving time, money and manpower, closer examination to identify specific lesions and stylish no matter experts in the diagnosis has made the application of neural networks in the diagnosis of abnormal situation and referral to a specialist is focus. In this paper we develop a neural network model to optimize the SVM, MLP, RBF for the diagnosis of major depressive disorder will be dealt with. The proposed scheme is based on the reactions of patients and healthy people to 45 related depression parameters such as depressed mood, loss of pleasure, Unknown disturbances in appetite (increase or decrease), Reduce energy, Feelings of guilt, Suicidal thoughts (such as a history of suicide or thinking about it), Worthless, Sleepless, Psychological distress, etc. The experimental results show that the detection errors of major depressive disorder have decreased to 3%, which reveals a high performance in primary diagnosis

Major Depressive, Neural Network, SVM, RBF, MLP