quotation:		[Copy]
		[Copy]

This Paper:Browse 4204 Download 2144
基于支持向量回归的现代气象雷达寿命预测
方睿,李享梅,涂爱琴
0 Fontlarge +\|Default\|Small
(成都信息工程学院网络工程学院，成都 610225;山东省气象局大气探测技术保障中心，济南 250031)

摘要:

为了准确地预测气象雷达使用寿命, 提高气象雷达运行可靠性、安全性和可维护性，提出了一种新的基于多元回归的气象雷达使用寿命预测框架。在该框架内，采用了支持向量回归(SVR)方法来求解气象雷达使用寿命的多元回归问题；提出了基于SVR的气象雷达特征参数选择FSSSVR（Feature Subset Selection SVR）算法去掉冗余和无效的特征参数。实验结果表明，基于SVR方法的预测算法能够准确地预测气象雷达的使用寿命，能够为雷达全寿命周期管理提供参考依据。

关键词: 气象雷达使用寿命预测支持向量回归全寿命周期管理

DOI：

基金项目:公益性行业（气象）科研专项资助项目(GYHY201006051)

Weather radar’s life prediction using support vector regression

FANG Rui,LI Xiang-mei,TU Ai-qin

()

Abstract:

In order to accurately predict the life of weather radar and improve its r eliability, security, maintainability, a novel framework for pr edicting weather radar′s life is proposed, which is based on multiple regressio n method.Within the framework, Support Vector Regression (SVR) is induced to sol ve the multivariate regression problem of weather radar′s life， and then, a fe ature subset selection SVR(FSSSVR) algorithm is developed to remove redundant an d ineffective components of characteristic parameters. Finally, simulation results demonstrate that SVRbased prediction algorithm is able to accurately forecast the life of weather radar, and provide reference for full lifecycle management of weather radar.

Key words: weather radar life prediction support vector regression life cycle management