Forecasting Study of Natural Gas Consumption by Combined Models Based on LASSO and WOA

As the impact of Russia-Ukraine war continues to expand, energy shortages appear in Europe. After Russia cut off Nord Stream 1 pipeline that transported natural gas Europe, most European countries experienced a crisis, severely affecting Germany. In order effectively predict consumption gas, this paper combines Least Absolute Shrinkage and Select Operator model with Whale Optimization Algorithm, uses NAR reconstruct phase space original time series, performs 5-step forward forecast. Use forecast German monthly dataset. Comparing results WOA-LASSO other five WOA-based hybrid models Cross-Validation based for prediction results, it is found has smallest MAPE each step prediction, numerical are between 8.273% 9.867%. Moreover, when comparing WOA conventional optimization scheme Cross-Validation, can obtain better hyperparameters, which enhance generalization performance accuracy model.