Traffic Status Prediction of Arterial Roads Based on the Deep Recurrent Q-Learning

Table 4

Experimental result table.

Group	Highest loss index	Lowest loss index	Maximum volatility	Training time (s)	Prediction time (s)	Total time (s)
1	0.240	0.082	0.0400	1080	100	1180
2	0.230	0.066	0.0385	1088	102	1190
3	0.195	0.070	0.0375	950	92	1042
4	0.970	0.096	0.4350	1086	101	1187
5	0.154	0.070	0.0435	1096	102	1198
6	0.250	0.080	0.0600	1139	105	1244
7	0.186	0.062	0.0510	1000	93	1093
8	0.195	0.080	0.0400	1033	96	1129
9	0.157	0.065	0.0440	1046	99	1145
10	0.238	0.070	0.0400	1013	94	1107
11	0.235	0.081	0.0440	1024	96	1120