Kernel-based clustering for short-term load forecasting

摘要

Precise short-term load forecasting is a very important task for power system operation. This paper presents three cascaded modules for short-term load forecasting: kernel-based clustering, decision tree and support vector regression. Kernel-based clustering classifies the input data into clusters efficiently. The data in every cluster have more similarity than in the original dataset. Decision tree is an effective way to decide which cluster the input data belong to. Support vector regression is used to predict daily load due to the advantages of structural risk, simple mathematical model and short training time. The concept of repetitious clustering and overlapping clusters are introduced to get better prediction precision. The effectiveness of the proposed method is demonstrated through comparison of the real load data with forecasted data.