基于最小二乘支持向量机的短期负荷预测模型  

Short-term Load Forecasting Model Based on Least Square Support Vector Machine

作　　者：张宁[1] 许承权[2] 薛小铃[1] 郑宗华[1] 

机构地区：[1]闽江学院物理学与电子工程系,福建福州350108 [2]闽江学院地理科学系,福建福州350108

出　　处：《现代电子技术》2010年第18期131-133,共3页Modern Electronics Technique

基　　金：福建省自然科学基金资助项目(2009J05102);福州市科技计划资助项目(2009-S-95);福建省教育厅A类科技项目(JA09193)

摘　　要：支持向量机(SVM)是近年来发展起来的机器学习的新方法,它较好地解决了小样本、非线性、高维数、局部极小点等实际问题。研究了支持向量机的拓展算法——最小二乘支持向量机(LSSVM),并将其应用于电力系统短期负荷时间序列预测。通过实例并与神经网络模型预测结果相比较表明,LSSVM模型的预测精度要明显高于神经网络模型,验证了LSS-VM模型可以很好地应用于短期负荷时间序列预测,并且具有较高的准确性与有效性,这为短期负荷预测提供了一个新的解决思路。Support vector machine （SVM） is a novel machine learning method, which is powerful for solving the problems characterized by small samples, nonlinearity, high dimension, local minima, and other practical problems. The extension algorithm of the support vector machine-least squares support vector machine（LSSVM） is studied and applied to the time series forecasting of short-term load in power system. Compared with the predicting outcomings of the neural network model through some examples, the result shows that the forecast accuracy of LSSVM model is higher than that of neural network model, and verifies LSSVM model can apply effectively to short-term load forecasting. This provides a new resolution for short-term load forecasting .

关 键 词：最小二乘支持向量机 神经网络 短期负荷预测 时间序列预测 

分 类 号：TN715.34[电子电信—电路与系统]