Adaptive Correction Forecasting Approach for Urban Traffic Flow Based on Fuzzy 𝑐-Mean Clustering and Advanced Neural Network

Table 2

The result of adding the correction mechanism.

ℎ+Δℎ	Monday 1		Wednesday 1		Sunday 1
	MSE	MAXE	MSE	MAXE	MSE	MAXE
ℎ+Δℎ=0	10.0456	43.1843	11.2106	45.2548	9.8952	42.1001
−0.1≤Δℎ≤0.1	9.9758	42.1285	10.2654	44.1657	9.7561	40.6548
−0.2≤Δℎ≤0.2	15.2648	59.2154	16.2299	61.2147	14.3215	57.6519
−0.3≤Δℎ≤0.3	99.2154	70.2165	105.2647	85.2594	99.0147	69.2589
ℎ+Δℎ	Monday 2		Wednesday 2		Sunday 2
	MSE	MAXE	MSE	MAXE	MSE	MAXE
ℎ+Δℎ=0	11.5248	40.1257	11.3017	47.0984	12.3014	48.2549
−0.1≤Δℎ≤0.1	10.4529	38.2489	10.3594	46.2813	11.2497	47.9523
−0.2≤Δℎ≤0.2	16.5489	58.2146	17.2016	64.0525	16.2018	58.2687
−0.3≤Δℎ≤0.3	103.4269	88.2159	106.2184	86.3468	101.4512	98.1264