Determination of Schizophrenia Using Electronic Nose via Support Vector Machine

Schizophrenia is a chronic brain disorder that is considered as the top mental illness in the Philippines. When schizophrenia is not prevented, especially on early stage more complications on the brain functionality is affected due to the deterioration of gray matter or brain tissues. Breath Analysis using electronic nose (e-nose) is a non-invasive and easy to use method of detection in providing information for many illnesses. This paper examines and measures the concentration of pentane, ammonia and other volatile organic compounds (VOCs) in determining schizophrenia using breath analysis. The preliminary performance of the electronic nose has been demonstrated on 10 breath samples from a subject that has schizophrenia than in healthy individual subject which has found out that the e-nose can discriminate the patterns of VOCs from these two groups. Confusion matrix was used to show the accuracy of the system in detecting schizophrenia. This study uses support vector machine for classification and achieving the accuracy of 80% of classifying schizophrenic and non-schizophrenic subjects.