Author/Authors :
Khosravanian, Asieh Department of Computer and Engineering and IT, Payame Noor University, Iran. , Ayat, Saeed Department of Computer and Engineering and IT, Payame Noor University, Iran
Abstract :
Introduction: Since human health is the issue of Medical Research, correct prediction of
results is of a high importance. This study applies probabilistic neural network (PNN) for
predicting coronary artery disease (CAD), because the PNN is stronger than other
methods.
Methods: In this descriptive-analytic study, The PNN method was implemented on 150
patients admitted to the Mazandaran Heart Center, sari. For designing the network, 80%
of the data were used for stage of network training, and the remained 20% were used for
stage of network testing. In order to implement the network, facilities and functions
existing in MATLAB 7.12.0 were used and simulation was conducted in a PC with
configurations of corei5 CPU, 2GHz processor, 4GB ram, under operating system of
Windows 7.
Results: After 5 times simulation and comparison of the models produced, sensitivity and specificity rates obtained were 1 and 1. In the end, model correctly categorized some
healthy subjects who did not need angiography and the treatment related to coronary
artery disease.
Conclusion: Due to the high specificity index, this model prevents side effects of
angiography in patients who don't need such treatments. Moreover, due to high
sensitivity, it can diagnose the patients who really need such diagnostic measures.
