Integrated Feature Selection of ARIMA with Computational Intelligence Approaches for Food Crop Price Prediction

Table 9

The results of MARS₁ for wheat price forecasting.

Variable selection results		BF	Coefficient
Variable	Relative importance (%)
Intercept			360.973
	100.0	BF1 = Max (0, yt−12 − 321.18)	0.765
		BF2 = Max (0, 321.81 − yt−12)	−1.272
	7.2	BF3 = Max (0, yt−6 − 178.32)	−0.185
		BF4 = Max (0, 178.32 − yt−6)	0.298
	3.2	BF5 = Max (0, yt−11 − 269.03)	−0.218