Feature Screening via Mutual Information Learning Based on Nonparametric Density Estimation

Table 2

Simulation result for Example 1.

		Method	Categories				MMMS
		PC-SIS		0.10	0.07	0.01	265.0 (137.3)
				0.64	0.69	0.45	45.5 (17.9)
				0.75	0.79	0.59	14.0 (8.2)
		MV-SIS	—	0.35	0.39	0.18	73.0 (85.8)
		MI-SIS	—	0.98	0.95	0.94	2.0 (0.7)
		PC-SIS		0.10	0.06	0.01	304.5 (126.3)
				0.97	0.96	0.93	2.0 (0.7)
				1.00	1.00	1.00	2.0 (0.0)
		MV-SIS	—	0.90	0.92	0.83	4.0 (1.5)
		MI-SIS	—	1.00	1.00	1.00	2.0 (0.0)
		PC-SIS		0.06	0.05	0.01	393.0 (194.0)
				0.55	0.59	0.30	127.0 (126.9)
				0.64	0.64	0.40	99.0 (22.4)
		MV-SIS	—	0.16	0.15	0.02	285.5 (96.3)
		MI-SIS	—	0.92	0.91	0.84	7.0 (4.4)
		PC-SIS		0.06	0.08	0.00	453.0 (97.0)
				0.96	0.95	0.92	2.0 (0.7)
				1.00	0.98	0.98	2.0 (0.0)
		MV-SIS	—	0.84	0.82	0.72	9.5 (4.5)
		MI-SIS	—	1.00	1.00	1.00	2.0 (0.0)

The user-specified number .