Driver and Path Detection through Time-Series Classification

Table 6

, precision, recall, accuracy, and training time evaluated by using MLP for different windows sizes.

Average window size	Driver	Driver recall	Driver precision	Overall precision	Overall recall	Accuracy	Training time (sec)
1 s; 14592 training segments	DF1	0,65	0,55	0,53	0,56	0,58	4320
	DM1	0,45	0,54
	DM2	0,49	0,6
	DM3	0,52	0,53
10 s; 1459 training segments	DF1	0,56	0,61	0,58	0,57	0,63	458
	DM1	0,58	0,53
	DM2	0,62	0,55
	DM3	0,55	0,59
30 s; 486 training segments	DF1	0,62	0,66	0,65	0,67	0,68	178
	DM1	0,64	0,69
	DM2	0,66	0,62
	DM3	0,68	0,71
60 s; 243 training segments	DF1	0,86	0,78	0,84	0,83	0,82	88
	DM1	0,83	0,81
	DM2	0,81	0,83
	DM3	0,86	0,88
Start&Stop; 7562 training segments	DF1	0,94	0,81	0,91	0,92	0,92	1740
	DM1	0,84	0,93
	DM2	0,92	0,97
	DM3	0,93	0,95