Electricity Load Forecasting in Smart Grids Using Support Vector Machine

摘要: One of the key issues in Smart Grid (SG) is accurate electric load forecasting. Energy generation and consumption have highly varying. Accurate forecasting can decrease fluctuating behavior between energy consumption. By knowing upcoming electricity consumption, we control extra generation. To solve this issue, proposed a model, which consists two-stage process; feature engineering classification. Feature selection extraction. combining Extreme Gradient Boosting (XGBoost) Decision Tree (DT) techniques, hybrid selector to minimize redundancy. Furthermore, Recursive Elimination (RFE) technique applied for dimension reduction improve selection. forecast load, Support Vector Machine (SVM) set tuned with three super parameters, i.e., kernel parameter, cost penalty, incentive loss function parameter. Electricity market data used our model. Weekly months ahead experiments are conducted by Forecasting performance assessed using RMSE MAPE their values 1.682 12.364. The simulation results show 98% accuracy.