Load Forecasting for Demand Side Management in Smart Grid using Non-Linear Machine Learning Technique
In recent times, as a result of the significant growth in population, the usage of energy consumptions is increasing rapidly. Analysis and design of the forecasting model for energy consumptions are challenging task because it is a complex and non-linear problem. Several methods have been proposed for forecasting the energy consumptions in smart grid, the non-linear relationship between the factors is not addressed. Therefore, there is a need for efficient, reliable and accurate forecasting methods to handle non-linearity for effective planning and management of energy consumption. In this paper, a novel hierarchical non-linear machine learning technique Multivariate Adaptive Regression Splines with Genetic Algorithm is proposed to manage the energy consumption demand in smart grid. Experimental results show that the proposed hierarchical approach is much more accurate for the forecasting of energy consumption in smart grid than other approaches such as Auto-Regressive Integrated Moving Average, Complex Neural Network and a Simple Regression Models. The evaluation of forecast accuracy measurement gives the least error value based on the performance metrics of the Mean Absolute Percentage Error and Root Mean Square Error are 0.0651 and 201.2381 respectively.