Table 1.

Classification accuracy (represented in percentage %) on different dataset.

	Drive	Big neuron	IJCNN1	Madelon
Our method	79.42	74.16	89.36	59.57
Standard RF	60.90	70.93	79.89	51.33
Robust node RF	63.79	70.99	78.92	50.53
Optimal RF	85.53	75.55	91.29	67.10

We show the accuracy on the training sample of 400 (DRIVE), 1,500 (Big Neuron), 300 (IJCNN1), and 400 (Madelon).