Driver Distraction Identification with an Ensemble of Convolutional Neural Networks

Table 2

Confusion matrix of our genetically weighted ensemble of classifiers on randomly selected test data.

		Predicted
		C0	C1	C2	C3	C4	C5	C6	C7	C8	C9
Actual	C0	95.34	0	0.33	0.65	0.11	0.43	0.43	0.87	0.11	1.74
	C1	0.31	96.63	1.23	0.31	0.92	0	0.31	0	0.31	0
	C2	0.29	3.23	96.48	0	0	0	0	0	0	0
	C3	2.02	0.61	0	96.15	0.81	0	0.20	0	0	0.20
	C4	0	0.33	0	4.90	94.77	0	0	0	0	0
	C5	4.26	0	0	0.33	0	95.08	0	0	0	0.33
	C6	0.74	0	0	0.25	0	0.74	98.01	0.25	0	0
	C7	3.65	0	0	0	0	0	0	95.35	0	1.00
	C8	3.79	0	0	0	0	0	1.38	0.34	92.76	1.72
	C9	1.40	0	0	0	0	0	0.47	0.31	0.16	97.67