Feature Selection for Intelligent Firefighting Robot Classification of Fire, Smoke, and Thermal Reflections Using Thermal Infrared Images

Table 5

Results of error, case, and performance at each feature combination in the behavioral solution set (Resu. means resubstitution and Cros. refers to cross-validation).

Index	Combination of features	Error (%)		Case			Performance (%)
		Resu.	Cros.	TP	FP	FN	Precision	Sensitivity	-measure	Accuracy
1	MNI, VAR, ENT, COR, SKE	6.90	6.89	10049	402	324	96.15	96.88	96.51	93.26
2	MNI, DIS, COR, SKE, STD	6.68	6.70	10069	459	247	95.64	97.61	96.62	93.45
3	MNI, ENT, COR, SKE, STD	6.76	6.75	10061	447	267	95.75	97.41	96.57	93.37
4	MNI, VAR, DIS, ENT, CON, COR, SKE, STD	6.98	6.90	9980	454	341	95.65	96.70	96.17	92.62
5	MNI, VAR, DIS, ENT, COR, UNI, SKE, STD	6.89	6.90	10037	453	285	95.68	97.24	96.45	93.15
6	MNI, VAR, DIS, ENT, COR, IDM, SKE, STD	6.77	6.79	10047	461	267	95.61	97.41	96.50	93.24
7	MNI, VAR, DIS, ENT, IDM, SKE, KUR, STD	6.96	6.94	10033	505	237	95.21	97.69	96.43	93.11
8	MNI, VAR, DIS, ENT, IND, COR, UNI, SKE, STD	6.95	6.99	10029	451	295	95.70	97.14	96.41	93.08
9	MNI, VAR, DIS, ENT, IND, COR, IDM, SKE, STD	6.88	6.90	10035	457	283	95.64	97.26	96.44	93.13
10	MNI, VAR, DIS, ENT, COR, IDM, SKE, KUR, STD	6.89	6.91	10036	478	261	95.45	97.47	96.45	93.14