The Predictive Method of Power Load Based on SVM

Abstract

For the problems which exist in power load prediction, such as nonlinear and high dimension recognition, support vector machine (SVM) is applied to power load prediction technology, based on the theory analysis of SVM and power load prediction. The predictive model of power load is established by SVM, which overcomes the problems to a great extent such as dimension disaster, over-learning, etc. The new thought of power load prediction is proposed which is more accurate, more intelligent and more humanized. The historical load data from a power grid company are used to establish the predictive model, Simulation is done by Matlab, and the predictive precision is improved greatly.

DOI : http://dx.doi.org/10.11591/telkomnika.v12i4.4834