Research Article
Integrated Feature Selection of ARIMA with Computational Intelligence Approaches for Food Crop Price Prediction
Table 4
The results of MARS1 for rice price forecasting.
| Variable selection results | BF | Coefficient | Variable | Relative importance (%) |
| Intercept | | | 542.484 | | 100.0 | BF1 = Max (0, yt−12 − 367) | 0.985 | BF2 = Max (0, 552.09 − yt−12) | −1.258 | | 13.1 | BF3 = Max (0, yt−11 − 356) | 1.539 | BF4 = Max (0, yt−11 − 472.48) | −1.880 | BF5 = Max (0, 585.95 − yt−11) | 0.524 | | 12.1 | BF6 = Max (0, yt−10 − 363) | −2.410 | BF7 = Max (0, yt−10 − 472.48) | 1.783 | BF8 = Max (0, yt−10 − 543.14) | 0.454 | BF9 = Max (0, 623 − yt−10) | −0.234 | BF10 = Max (0, yt−10 − 623) | 0.346 |
|
|