Support vector regression with asymmetric loss for optimal electric load forecasting

In energy demand forecasting, the objective function is often symmetric, implying that over-prediction errors and under-prediction have same consequences. practice, these two types of generally incur very different costs. To accommodate this, we propose a machine learning algorithm with cost-oriented asymmetric loss in training procedure. Specifically, develop new support vector regression incorporating linear-linear cost insensitivity parameter for sufficient fitting. The electric load data from state New South Wales Australia used to show superiority our proposed framework. Compared basic regression, framework multi-step forecasting results daily economic reduction ranging 42.19% 57.39%, depending on actual ratio errors.