Power Load Forecasting Based on the Locally Weighted Support Vector Machines

Accurate forecasting of electricity load has been one of the most important issues in the electricity industry. Modern data mining methods have played a crucial role in forecasting electricity load. Support Vector Machines (SVMs) have been successfully employed to solve nonlinear regression and time series problems. In view of the recent load have greater impact on the predicted results, the paper improves traditional support vector machine, and proposes a Locally Weighted Support Vector Machines (LW-SVMs). The methodology is applied to the case of load forecasting in Inner Mongolia of China. The results shows that power load forecasting using Locally Weighted Support Vector Machine have much more accurate result than traditional Support Vector Machine.