Integrated Framework for Prognosis of Coronary Artery Disease

Coronary Artery Disease, commonly known as Heart Disease, has resulted in casualties all over the world. Currently, diagnosis of these  conditions mainly relies on experience of doctors and partly on few decision support systems. Most decision support systems built for  this domain have limitations. The major limitation is requirement of huge quantity of historical conditions data and appropriate related diagnosis labels. Proposed framework addresses this limitation by providing diagnosis  comparable with existing systems while only taking available data in absence of labels.  This is achieved by forming a novel framework using Self Organizing Networks and  Learning Vector Quantization. Here, competitive learning paradigm of soft  computing is primary focus. Proposed framework is compared with four well established  existing systems namely Support Vector Machine, Decision Tree, Random Forest and Multi Layer Perceptron. After comparison and analysis it has been proved that proposed framework gives comparable results even without supplying labeled data.In future this framework can be extended to variegated applications in various domains.