Optimal Hybrid Modeling Scheme for Electricity Price Prediction Using Ensemble Learning

One of the most crucial aspects of decision-making for electricity market participants to develop rational competitive strategies is accurate energy price forecast. An important task in energy conservation is the prediction of energy prices and use. Support vector regression performs well when dealing with non-linear data issue with data regression It has been employed frequently in recent years to forecast building energy costs and usage. We draw the conclusion that the relationship between the cost of lighting energy, its consumption, and its influencing factors is non-linear based on historical data. For energy trading organizations, predicting energy prices and usage is a crucial task. As the accuracy of the prediction directly affects the company's profit, it should be as precise as possible. One of the most crucial tasks in the management and planning of power systems is the forecasting of electrical loads, energy prices, and consumption. Yet occasionally, we have to find a solution to this issue without access to enough and trustworthy historical data. The support vector regression with radial basis function was used to construct an accurate prediction model of lighting energy price and usage. The forecast findings show that support vector regression has greater prediction accuracy than neural networks. The prediction model can project building energy costs and consumption on an hourly basis and evaluate the effects of energy management strategies for office buildings. Once the base models have been selected, the next step is to construct the ensemble. The most common approach is to use a weighted average of the predictions of the base models. The weights can be determined using various techniques, such as cross-validation or gradient-based optimization.