An Improved Generalized-Trend-Diffusion-Based Data Imputation for Steel Industry

Table 3

Imputation accuracies with different methods (coke oven’s BFG consumption).

Missing number	Methods	Group 1			Group 2			Group 3
		RMSE	NRMSE	MAPE	RMSE	NRMSE	MAPE	RMSE	NRMSE	MAPE
	Regression	16.73	0.0300	7.90	10.33	0.0185	3.56	9.37	0.0167	3.24
	EM	9.95	0.0178	5.43	7.25	0.0130	4.05	3.72	0.0066	2.40
3 points	SPLINE	18.77	0.0336	7.73	16.81	0.0301	7.79	22.64	0.0402	9.49
	GTD	15.83	0.0283	8.30	15.82	0.0283	7.47	15.67	0.0278	7.49
	iGTD	13.94	0.0250	6.76	13.76	0.0246	6.55	13.74	0.0244	7.30
	Regression	16.66	0.0260	7.81	11.10	0.0171	4.18	10.82	0.0168	4.49
	EM	10.32	0.0161	5.47	6.72	0.0104	3.80	6.71	0.0104	3.16
4 points	SPLINE	15.77	0.0246	6.50	38.26	0.0591	9.33	14.97	0.0231	6.37
	GTD	15.93	0.0249	7.44	19.75	0.0305	10.55	29.51	0.0459	13.58
	iGTD	13.81	0.0216	6.07	19.36	0.0299	9.39	13.02	0.0203	6.51
	Regression	13.24	0.0186	5.73	13.24	0.0184	5.55	15.98	0.0227	7.78
	EM	10.34	0.0145	5.31	7.52	0.0104	4.20	10.78	0.0153	5.27
5 points	SPLINE	16.90	0.0237	9.06	21.57	0.0300	8.51	13.85	0.0196	6.20
	GTD	15.49	0.0217	8.33	22.83	0.0317	11.00	20.23	0.0287	11.48
	iGTD	11.81	0.0166	5.45	19.84	0.0276	9.02	19.27	0.0273	9.45