首页 | 本学科首页   官方微博 | 高级检索  
     检索      

一种新的晶化过程电特性预测方法
引用本文:王晓兰,康蕾,王海洋.一种新的晶化过程电特性预测方法[J].甘肃科学学报,2009,21(4):35-38.
作者姓名:王晓兰  康蕾  王海洋
作者单位:1. 兰州理工大学,电气工程与信息工程学院,甘肃,兰州,730050;甘肃省有色金属新材料省部共建国家重点实验室,甘肃,兰州,730050
2. 兰州理工大学,电气工程与信息工程学院,甘肃,兰州,730050
基金项目:甘肃省自然科学基金项目,甘肃省有色金属新材料省部共建国家重点实验室开放基金项目 
摘    要:金属玻璃具有许多独特性能,有着广泛的应用前景,晶化过程对金属玻璃的特性有重要影响.因此提出了一种基于小波变换和AR-ILSSVM金属玻璃晶化过程电特性预测方法.首先利用小波变换将时间序列分解为高频序列和低频序列,利用AR模型预测高频序列,利用ILSSVM方法预测低频序列,预测结果为上述预测的合成.仿真试验表明,这种方法可显著提高时间序列的预测精度.

关 键 词:小波变换  最小二乘支持向量机  增量式算法  金属玻璃晶化  时间序列预测

A New Method of Forecasting about Electric Characteristic of Crystallization
WANG Xiao-lan,KANG Lei,WANG Hai-yang.A New Method of Forecasting about Electric Characteristic of Crystallization[J].Journal of Gansu Sciences,2009,21(4):35-38.
Authors:WANG Xiao-lan  KANG Lei  WANG Hai-yang
Institution:WANG Xiao-lan1,2,KANG Lei1,WANG Hai-yang1(1.School of Electrical and Informational Engineering,Lanzhou University of Science and Technology,Lanzhou 730050,China,2.State Key Laboratory of Non-ferrous Metal Materials,China.)
Abstract:The metallic glasses have been widely used in engineering due to their special properties.The process of crystallization has great affects on the properties of metallic glasses.An algorithm for forecasting electrical characteristics of metallic glasses based on wavelet transform and incremental LSSVM was proposed.By wavelet transform,the time series are decomposed into the low frequency series and several high frequency series.The AR models are used to forecast the high frequency series,and the incremental ...
Keywords:wavelet transform  least square support vector machine  incremental algorithm  crystallization of metallic glass  time series forecasting  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号