[Retracted] Wheat Futures Prices Prediction in China: A Hybrid Approach

Table 6

The Diebold–Mariano future prediction accuracy test.

∶ The model and random walk under contemplation execute similarly fighting fit
∶ RW outperforms the model
	Artificial neural network	ARIMA (1, 1, 1)	ARIMA (1, 1, 2)	RW	Conventional model
MAE	0.4163	0.4975	0.4986	0.4948	0.4301
RMSE	0.7999	0.7673	0.7678	0.7627	0.7766

Note., , and specify statistical significance at the 10%, 5%, and 1% levels, respectively.