Table 6. Comparison of the performance of various algorithms for dataset 1, 2, and 3.

The proposed method showed the best performance for all three datasets.

Dataset	Algorithm	Weighted average
		Accuracy	Precision	Recall	F-Measure	ROC area
1	Naïve Bayes	0.537	0.551	0.537	0.504	0.581
	SVM	0.580	0.580	0.580	0.579	0.580
	ANN	0.570	0.570	0.570	0.570	0.603
	PART	0.742	0.742	0.742	0.742	0.842
	Proposed method	0.902	0.905	0.902	0.902	0.954
2	Naïve Bayes	0.547	0.567	0.547	0.512	0.585
	SVM	0.562	0.564	0.562	0.559	0.562
	ANN	0.567	0.567	0.567	0.567	0.597
	PART	0.713	0.723	0.713	0.710	0.812
	Proposed method	0.898	0.899	0.898	0.898	0.953
3	Naïve Bayes	0.549	0.567	0.549	0.518	0.597
	SVM	0.563	0.571	0.563	0.549	0.563
	ANN	0.570	0.570	0.570	0.570	0.601
	PART	0.744	0.746	0.744	0.743	0.850
	Proposed method	0.916	0.916	0.916	0.916	0.965
4	Naïve Bayes	0.529	0.533	0.529	0.515	0.555
	SVM	0.552	0.552	0.552	0.551	0.552
	ANN	0.535	0.537	0.535	0.528	0.565
	PART	0.628	0.628	0.628	0.628	0.704
	Proposed method	0.783	0.783	0.783	0.782	0.861
5	Naïve Bayes	0.540	0.560	0.540	0.499	0.577
	SVM	0.556	0.580	0.556	0.522	0.556
	ANN	0.559	0.559	0.559	0.559	0.587
	PART	0.642	0.644	0.642	0.640	0.718
	Proposed method	0.772	0.773	0.772	0.772	0.851
6	Naïve Bayes	0.535	0.552	0.535	0.494	0.571
	SVM	0.555	0.583	0.555	0.515	0.555
	ANN	0.565	0.566	0.565	0.565	0.591
	PART	0.662	0.662	0.662	0.662	0.752
	Proposed method	0.786	0.786	0.786	0.786	0.865