ILRC: a hybrid biomarker discovery algorithm based on improved L1 regularization and clustering in microarray data

Table 1 Classification accuracy of different sampling methods on unbalanced datasets

	Datasets	SVM	GB	DT	NN	KNN
Origin datasets	ALL1	1.0000	1.0000	0.9843	1.0000	1.0000
	ALL3	0.8080	0.8640	0.7680	0.8240	0.8080
	ALL4	0.9567	0.9053	0.8053	0.9129	0.8930
	DLBCL	0.9875	0.9750	0.8558	0.9875	0.9750
	Myeloma	0.9018	0.8965	0.7114	0.8842	0.8382
Over sampling	ALL1	1.0000	1.0000	1.0000	1.0000	1.0000
	ALL3	0.8261	0.8271	0.8371	0.9307	0.7326
	ALL4	0.9704	0.9254	0.8644	1.0000	0.8729
	DLBCL	0.9482	0.9312	0.9141	0.9739	0.8790
	Myeloma	0.8793	0.8613	0.8359	0.8687	0.8100
Combined sampling	ALL1	1.0000	1.0000	1.0000	1.0000	1.0000
	ALL3	0.8766	0.7926	0.7918	0.9456	0.7616
	ALL4	0.9704	0.9259	0.7610	0.9630	0.8667
	DLBCL	0.9913	0.9913	0.8877	1.0000	0.9830
	Myeloma	0.8943	0.8650	0.8250	0.8725	0.8100

ISSN: 1471-2105