Research Article
A Crude Oil Spot Price Forecasting Method Incorporating Quadratic Decomposition and Residual Forecasting
Table 9
Characteristics of different models.
| | | Decomposition technique | Two-layer decomposition technique |
| | VAR (model 1) | | | | BPNN (model 2) | | | | RF (model 3) | | | | ARIMA (model 4) | | | | ELM (model 5) | | | | XGBoost (model 6) | | | | SVR (model 7) | | | | SOA-SVR (model 8) | | | | VMD-SOA-SVR (model 9) | √ | | | VMD-CEEMDAN-SOA-SVR (model 10) | √ | √ | | CEEMDAN-SOA-SVR (model 11) | √ | | | CEEMDAN- RES.-SOA-SVR (model 12) | √ | | | CEEMDAN -RES.- VMD -SOA-SVR (model 13) | √ | √ |
|
|
Ps: model 12 and model 13 are decompositions that include the residual term after the first decomposition.
|
