Journals Proceedings

International Journal of Artificial Intelligence and Neural Networks

An Echocardiogram Diagnosis System based on LDA and Fuzzy C-Mean Unsupervised Classifier

Author(s) : MURAT KARABATAK

Abstract

Today, many people are at risk of heart attack. Echocardiogram imaging techniques, looking at the eye doctors to view patients with heart attack risk is difficult to determine. In this case, the possibility of making an incorrect diagnosis by doctors rises. Recently, many computer-based decision support systems have been developed to decrease wrong diagnosis possibility in the field of echocardiogram. The main purpose of these decision support systems designed to help expert doctors in the field of medical diagnostics. In this study, an echocardiogram diagnostic (ED) based on Linear Discriminant Analysis (LDA) and Fuzzy C-Mean (FCM) clustering algorithm (LDA-FCM-ED) system is presented. This echocardiogram diagnostic based on Linear Discriminant Analysis and Fuzzy C-Mean clustering algorithm system is composed of two stages. First step is the feature extraction and feature reduction stage by using LDA and second step is clustering stage by using FCM unsupervised classifier. In stage of feature reduction, LDA is used to reduce the attributes from 12 to 4. In clustering stage, features obtained in first stage are fed into the FCM unsupervised classifier. The correct diagnosis performance of the LDA-FCM-ED clustering system for diagnosis the echocardiogram is calculated by using classification accuracy, sensitivity and specificity values respectively. The classification accuracy of this LDA-FCM clustering system for diagnosis of the echocardiogram (LDAFCM- ED) was obtained 93.44 %. This result is better than FCM and other existing unsupervised classifier methods

No fo Author(s) : 1
Page(s) : 6 - 10
Electronic ISSN : 2250-3749
Volume 5 : Issue 2
Views : 449   |   Download(s) : 115