支持向量机及其在用电量时间序列预测中的应用  

Application of Support Vector Machine in Time Series Prediction of Electric Consumption

作　　者：王绪玲 王炜立[2] 姚力文[2] 

机构地区：[1]江西省景德镇市商业职工中等专业学校,江西景德镇330029 [2]南昌大学计算机系,江西南昌330031

出　　处：《计算机与现代化》2008年第10期130-133,共4页Computer and Modernization

摘　　要：支持向量机在解决非线性及高维模式识别问题中表现出许多特有的优势。由于现实世界中大量数据的采集与时间相关,数据具有时间上的关联性,从而时间序列预测成为人们更感兴趣同时也是更富挑战性的工作。本文探讨了支持向量机对混沌时间序列的预测能力,推导出在用电量时间序列预测中的模型,并进行求解。可以看出将支持向量机理论和方法应用于电流量时间序列预测中具有理论和实际意义。The advantage of SVM is to solve the non-linear and pattern recognition with high dimension. Concerning mass data gathering is related to time due to the correlation between data and time in real world, the time series prediction is interesting and challenging. This paper discusses the SVM＇ s application to predict chaotic time series, constructs time series prediction model in electric consumption and performs derivation. It is obvious that application of SVM in time series prediction of electric current will be of great significance.

关 键 词：支持向量机 时间序列预测 电流量 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]