Application of decision tree method in the diagnosis of neuropsychiatric diseases

Abstract

In this paper, the Electroencephalogram (EEG) and Functional Magnetic Resonance Imaging (FMRI) parameters along with physical, cognitive and psychological parameters altogether used in the detection and diagnosis of five neuropsychiatric diseases. The diseases are considered for analysis and diagnosis are Attention Deficit Hyperactivity Disorder (ADHD), Dementia, Mood Disorder (MD), Obsessive-Compulsive Disorder (OCD) and Schizophrenia (SZ). The detection and diagnosis of disease depends upon the different parameters. In this work we are analyzing thirty eight parameters (five category) using C5.0 algorithm to know the importance and contribution of parameters in the diagnosis. The formation of decision tree based on C5.0 algorithm using Clementine tool is also verified using the manual calculation to know the important parameters at different levels in the tree. The decision tree structure gives doctors easiest way to analysis and diagnoses diseases based on important parameters. The results of C5.0 algorithm is also compared with our previous work i.e., Rule-based and Case-based reasoning model in the diagnosis of neuropsychiatric diseases. The comparative shows the accuracy of each model. © 2014 IEEE.