Table 1.

Misclassification rate of the methods on the data sets with added non-informative features from model 1

Features	kNN	BkNN	RkNN	MFS	RF	SVM	ESkNN
20	0.050	0.047	0.046	0.048	0.052	0.043	0.044
$20+50$	0.063	0.058	0.062	0.061	0.055	0.055	0.047
$20+100$	0.076	0.067	0.071	0.066	0.066	0.057	0.046
$20+200$	0.114	0.104	0.089	0.084	0.063	0.065	0.046
$20+500$	0.146	0.127	0.142	0.112	0.062	0.084	0.046

The first column shows the number of non-informative features added to the data set. Results of the best performing method are highlighted in italics. The value of $w=1$