Container Throughput Forecasting of Tianjin-Hebei Port Group Based on Grey Combination Model

Table 5

Comparison of BP neural network and the original data.

Year	Tian. original norm.	Tang. original norm.	Qin. original norm.	Huang. original norm.	Total original data norm.	Original data (unit: 10,000 TEU)	Total after BPNN training	Total data antinorm. (unit: 10,000 TEU)	Error
2012	0.01	0.01	0.01	0.01	0.01	1320	0.018	1320.00	0.00%
2013	0.10	0.10	0.15	0.19	0.11	1436	0.124	1446.95	0.76%
2014	0.24	0.22	0.20	0.31	0.24	1589	0.232	1576.69	−0.77%
2015	0.24	0.36	0.45	0.58	0.30	1663	0.331	1696.50	2.01%
2016	0.30	0.49	0.51	0.72	0.38	1758	0.396	1773.62	0.89%
2017	0.37	0.68	0.62	0.79	0.48	1881	0.442	1829.75	−2.72%
2018	0.49	0.82	0.67	0.89	0.60	2027	0.597	2016.10	−0.54%
2019	0.65	0.82	0.78	0.67	0.70	2143	0.659	2089.84	−2.48%
2020	0.79	0.88	0.78	0.90	0.82	2282	0.801	2261.02	−0.92%