A Parallel SVR Model for Short Term Load Forecasting Based on Windows Azure Platform

Short term load forecasting (STLF) is an important process in electric power operation and control system. Support Vector Regression (SVR) is proved to be a successful application in STLF, and can get great accuracy and efficiency compared to other STLF models. However, when deal large scale sample size, SVR is poor on the performance. With the development of cloud computing, it is changing people´s life in more and more areas. Windows Azure Platform is a cloud computing platform developed by Microsoft. It can easily scale up or down to get compute or storage resource according to requirements. Take into account the advantage and convenience, we propose a parallel SVR model based on Windows Azure Platform to solve the large scale dataset problem of SVR. This model is verified with ENUN standard dataset, the results shows that the model of SVR based on Windows Azure Platform has apparently improvement on efficiency than standard SVR model.