การเปรียบเทียบวิธีการจัดกลุ่มกรณีข้อมูลมีการแจกแจงเสถียรปกติ การแจกแจงโคชี และการแจกแจงเลวี่ (THE COMPARISON OF CLASSIFICATION WITH STABLE-NORMAL CAUCHY AND LE’VY DISTRIBUTIONS)

Abstract

The proposed of this research aimed is to compare the efficiency of support vector machine (SVM) and K-nearest neighbor (K-NN) when data are distributed as Stable-normal, Cauchy and Le’vy distributions. Data employed in this research were generated into two sets, consisting of training set and test set with the sample sizes 100, 300 and 500 for the binary classification. The ratios of training sets: test sets are 90:10, 80:20, 70:30, 60:40 and 50:50. In each situation, the data are simulated with Monte Carlo technique and repeated 1,000 times. The average percentage of correct classification is used as criterion for comparison. The results found that the SVM method produces higher percentages of correct classification than K-NN method in all situations under study.