Performance Comparison of Rule Generation Method Substractive Clustering and Fuzzy C-Means Clustering on Sugeno's Inference for Stroke Risk Detection

Abstract

- Fuzzy Inference is one method that cansolve the problem of uncertainty in a decision-makingor classification well. In inference, fuzzy rules thatrepresent the need of expert knowledge in the relevantfields, so that the classification given decision or beappropriate expert knowledge. However there are timeswhen experts are less able to represent the rules of theappropriate knowledge or knowledge that there is needof too many rules, so we need a method that cangenerate rules based on the data given expert.At issue troke s disease risk detection, it also occursbecause of the research that has been done by taking thedirect rule of experts, it turns out less than the maximumaccuracy, still 82.89%. Substractive methodsClustering and Fuzzy C-Means (FCM) could generaterules by grouping algorithm, in which the existingtraining data are grouped in common and the rules ofthe group raised. Differences in the two methods are indetermining the center of the cluster and assign eachincoming data which groups.Based on research that has been done, substractiveaverage Clustering membrika better accuracy is84.46%, while 73.81% FCM. However, in theprocessing time FCM faster at 16.75 seconds to give anaverage processing time of 13:02 seconds.