Wavelet Packet Transform Based Driver Distraction Level Classification Using EEG

Table 2

Classification accuracy of different classifiers for different wavelets over 4 distraction levels for both two features.

	Neutral		Low		Medium		High		Average
	Centroid	PSD	Centroid	PSD	Centroid	PSD	Centroid	PSD	Centroid	PSD
Fuzzy
db4	71.17	77.73	67.08	79.21	79.32	84.80	64.87	70.99	70.61	78.18
db8	64.17	73.24	77.29	89.90	75.99	71.73	63.61	79.41	70.27	78.57
sym8	62.51	72.79	82.61	91.99	81.01	70.60	63.65	79.21	72.45	79.21
coif5	65.36	74.94	63.19	66.30	76.81	78.22	62.68	63.55	67.01	70.75
PNN
db4	62.39	58.13	58.14	73.56	65.88	83.20	31.19	63.57	54.40	69.62
db8	52.64	50.31	72.92	90.79	57.56	64.26	25.92	75.48	52.26	70.21
sym8	25.00	49.29	69.22	91.29	77.50	63.31	53.91	79.22	56.41	70.78
coif5	55.42	68.52	50.93	34.01	58.50	60.84	24.80	25.42	47.41	47.20
KNN
db4	44.29	72.33	34.93	61.69	75.79	77.75	55.26	45.49	52.57	64.32
db8	25.31	66.34	60.25	90.13	71.54	46.35	52.98	65.04	52.52	66.97
sym8	50.70	65.53	78.55	91.29	69.00	44.66	28.20	70.91	56.61	68.10
coif5	31.29	52.45	24.74	56.91	70.57	73.49	50.74	53.00	44.34	58.97

Bold values refers to maximum average values.