Performance Analysis of Fuzzy processor for a Healthcare Application-Diabetic Epilepsy Risk Classifier

A diabetic is a chronic disease. In worldwide 422 million people were affected. The burden of diabetics is increasing, specifically in developing countries. It produces one sort of disorder in which cluster of nerve cells in the brain function abnormally. So it is necessary to identify the epilepsy risk level accurately for the treatment. In this paper we propose a performance evaluation of VLSI based diabetic epilepsy risk level classifier using Xilinx ise 12.2 and plan ahead 12.2 software tool, and the optimized output were obtained using Neural network tool, Singular Value Decomposition. Epilepsy classifier fuzzy processor has been checked in both the environment like windows and Open Source environment (FOSS-Free and Open Source Software). The Xilinx ise 9.1i is used for the open source environment. Electroencephalogram (EEG) and Cerebral Blood Flow (CBF) level are used as input parameter. Confusion matrix has been tabulated and sensitivity, specificity, prevalence, positive predictive value and negative predictive value parameters were obtained. Higher degree of classification obtained in Very low, Medium levels. Singular value Decomposition (SVD) is used for the dimensionality reduction. area, power, delay analysis were performed with the resources utilization. The comparison between the VLSI based conventional diabetic classifier and synthesized fuzzy classifier were tabulated. Our results shows that proposed fuzzy processor output closely follows the previous Mat lab results in all the linguistic levels.