Table 8. Prediction performance for the Kansai dataset.

	RMSE
Node i	MA(18)	ES(0.05)	NN+BU	NN+MinT	NN+SR(0.4, 1.2)
Root	13.88	13.84	13.60 ± 0.68	12.47 ± 0.30	12.20 ± 0.39
2	2.57	2.56	2.58 ± 0.09	2.43 ± 0.05	2.37 ± 0.04
3	12.78	12.79	12.56 ± 0.69	11.57 ± 0.30	11.14 ± 0.41
4	1.90	1.90	1.83 ± 0.04	1.81 ± 0.04	1.68 ± 0.03
Mid-level	5.75	5.75	5.66 ± 0.12	5.27 ± 0.11	5.06 ± 0.07
5	0.73	0.74	0.77 ± 0.03	0.72 ± 0.02	0.79 ± 0.02
6	1.80	1.82	1.87 ± 0.07	1.77 ± 0.04	1.83 ± 0.04
7	1.35	1.36	1.33 ± 0.07	1.35 ± 0.05	1.22 ± 0.04
8	11.31	11.34	11.29 ± 0.66	10.38 ± 0.27	10.02 ± 0.39
9	2.71	2.69	2.62 ± 0.14	2.60 ± 0.12	2.43 ± 0.10
10	1.50	1.49	1.48 ± 0.07	1.45 ± 0.06	1.43 ± 0.06
11	1.16	1.14	1.14 ± 0.04	1.12 ± 0.04	1.03 ± 0.03
12	0.82	0.82	0.79 ± 0.02	0.81 ± 0.01	0.78 ± 0.01
13	0.99	0.99	0.96 ± 0.03	0.96 ± 0.03	0.95 ± 0.02
Bottom-level	2.49	2.49	2.47 ± 0.04	2.35 ± 0.04	2.28 ± 0.02
Average	4.12	4.11	4.06 ± 0.08	3.80 ± 0.07	3.68 ± 0.05