A Novel Robust Fuzzy Rough Set Model for Feature Selection

Table 3

Comparison of classification performance of different FRS models on CART in original and noisy data.

Datasets	Noise level (%)	FRS	β-PFRS	K-trimmed FRS	K-means FRS	K-median FRS	SFRS	RS-FRS
Glass	0	0.478 ± 0.031	0.436 ± 0.027	0.454 ± 0.039	0.458 ± 0.045	0.455 ± 0.025	0.478 ± 0.031	0.478 ± 0.031
	5	0.430 ± 0.025	0.434 ± 0.026	0.433 ± 0.010	0.436 ± 0.004	0.434 ± 0.049	0.463 ± 0.069	0.465 ± 0.060
	10	0.429 ± 0.069	0.431 ± 0.014	0.432 ± 0.025	0.431 ± 0.051	0.430 ± 0.052	0.454 ± 0.010	0.460 ± 0.025
Wine	0	0.909 ± 0.068	0.909 ± 0.068	0.908 ± 0.081	0.908 ± 0.081	0.908 ± 0.081	0.908 ± 0.081	0.909 ± 0.068
	5	0.859 ± 0.059	0.903 ± 0.092	0.901 ± 0.026	0.860 ± 0.075	0.899 ± 0.064	0.902 ± 0.072	0.903 ± 0.083
	10	0.819 ± 0.076	0.857 ± 0.037	0.897 ± 0.087	0.859 ± 0.041	0.893 ± 0.056	0.896 ± 0.054	0.902 ± 0.107
Heart	0	0.707 ± 0.019	0.759 ± 0.099	0.752 ± 0.089	0.707 ± 0.019	0.707 ± 0.019	0.707 ± 0.124	0.752 ± 0.052
	5	0.694 ± 0.054	0.737 ± 0.091	0.707 ± 0.019	0.702 ± 0.047	0.704 ± 0.010	0.705 ± 0.055	0.750 ± 0.049
	10	0.659 ± 0.045	0.728 ± 0.043	0.693 ± 0.011	0.682 ± 0.044	0.693 ± 0.031	0.704 ± 0.023	0.748 ± 0.076
Segment	0	0.947 ± 0.029	0.946 ± 0.039	0.945 ± 0.035	0.945 ± 0.035	0.945 ± 0.035	0.947 ± 0.029	0.947 ± 0.029
	5	0.941 ± 0.031	0.953 ± 0.027	0.942 ± 0.041	0.943 ± 0.034	0.942 ± 0.027	0.945 ± 0.083	0.946 ± 0.082
	10	0.938 ± 0.036	0.938 ± 0.030	0.940 ± 0.032	0.941 ± 0.035	0.941 ± 0.038	0.942 ± 0.064	0.944 ± 0.033
Hepatitis	0	0.800 ± 0.063	0.800 ± 0.063	0.800 ± 0.063	0.800 ± 0.063	0.800 ± 0.063	0.800 ± 0.063	0.800 ± 0.063
	5	0.762 ± 0.093	0.761 ± 0.103	0.770 ± 0.130	0.753 ± 0.081	0.755 ± 0.101	0.796 ± 0.039	0.798 ± 0.080
	10	0.723 ± 0.126	0.752 ± 0.079	0.732 ± 0.021	0.751 ± 0.104	0.728 ± 0.120	0.781 ± 0.041	0.790 ± 0.105
ICU	0	0.799 ± 0.085	0.783 ± 0.036	0.783 ± 0.036	0.783 ± 0.036	0.783 ± 0.036	0.799 ± 0.085	0.799 ± 0.085
	5	0.760 ± 0.062	0.764 ± 0.024	0.763 ± 0.022	0.773 ± 0.045	0.761 ± 0.019	0.762 ± 0.077	0.786 ± 0.091
	10	0.751 ± 0.013	0.755 ± 0.088	0.754 ± 0.096	0.755 ± 0.019	0.754 ± 0.049	0.752 ± 0.082	0.761 ± 0.114
German	0	0.716 ± 0.037	0.704 ± 0.035	0.713 ± 0.045	0.713 ± 0.032	0.713 ± 0.032	0.712 ± 0.046	0.716 ± 0.037
	5	0.716 ± 0.053	0.700 ± 0.031	0.711 ± 0.039	0.712 ± 0.034	0.710 ± 0.038	0.710 ± 0.071	0.713 ± 0.064
	10	0.707 ± 0.033	0.700 ± 0.061	0.709 ± 0.031	0.708 ± 0.042	0.709 ± 0.032	0.710 ± 0.079	0.709 ± 0.088
Soy	0	0.915 ± 0.111	0.915 ± 0.111	0.915 ± 0.111	0.915 ± 0.111	0.915 ± 0.111	0.915 ± 0.111	0.915 ± 0.111
	5	0.810 ± 0.067	0.883 ± 0.077	0.879 ± 0.004	0.891 ± 0.067	0.878 ± 0.058	0.894 ± 0.068	0.893 ± 0.108
	10	0.790 ± 0.039	0.810 ± 0.051	0.806 ± 0.058	0.808 ± 0.013	0.806 ± 0.025	0.821 ± 0.022	0.825 ± 0.024
Horse	0	0.855 ± 0.046	0.862 ± 0.049	0.857 ± 0.050	0.854 ± 0.047	0.857 ± 0.050	0.856 ± 0.052	0.855 ± 0.046
	5	0.846 ± 0.067	0.860 ± 0.053	0.854 ± 0.060	0.853 ± 0.050	0.852 ± 0.047	0.855 ± 0.013	0.855 ± 0.103
	10	0.841 ± 0.047	0.848 ± 0.055	0.848 ± 0.038	0.851 ± 0.033	0.853 ± 0.040	0.853 ± 0.051	0.854 ± 0.027
WDBC	0	0.907 ± 0.037	0.931 ± 0.046	0.907 ± 0.042	0.926 ± 0.044	0.916 ± 0.038	0.907 ± 0.037	0.928 ± 0.029
	5	0.903 ± 0.084	0.924 ± 0.031	0.903 ± 0.041	0.909 ± 0.033	0.914 ± 0.034	0.905 ± 0.102	0.927 ± 0.057
	10	0.895 ± 0.058	0.923 ± 0.022	0.905 ± 0.045	0.907 ± 0.032	0.896 ± 0.036	0.904 ± 0.046	0.924 ± 0.037
WPBC	0	0.660 ± 0.115	0.703 ± 0.128	0.645 ± 0.109	0.660 ± 0.134	0.634 ± 0.159	0.666 ± 0.090	0.660 ± 0.092
	5	0.630 ± 0.118	0.702 ± 0.112	0.640 ± 0.105	0.656 ± 0.153	0.631 ± 0.109	0.666 ± 0.086	0.659 ± 0.067
	10	0.620 ± 0.110	0.659 ± 0.104	0.629 ± 0.150	0.640 ± 0.135	0.629 ± 0.139	0.664 ± 0.053	0.659 ± 0.035
Sonar	0	0.668 ± 0.065	0.686 ± 0.122	0.688 ± 0.115	0.659 ± 0.066	0.659 ± 0.066	0.686 ± 0.065	0.668 ± 0.081
	5	0.645 ± 0.035	0.678 ± 0.099	0.654 ± 0.115	0.644 ± 0.035	0.643 ± 0.047	0.659 ± 0.012	0.666 ± 0.109
	10	0.634 ± 0.105	0.659 ± 0.106	0.647 ± 0.110	0.638 ± 0.088	0.636 ± 0.095	0.658 ± 0.025	0.665 ± 0.053
Average		0.755 ± 0.061	0.772 ± 0.063	0.764 ± 0.060	0.762 ± 0.056	0.761 ± 0.056	0.772 ± 0.059	0.779 ± 0.067