Wavelet compressibility of dense linear systems arising from numerical solution of integral equations

被引:0
|
作者
Li, Y [1 ]
Yan, Y [1 ]
Devel, M [1 ]
Langlet, R [1 ]
Song, G [1 ]
机构
[1] Xidian Univ, Sch Sci, Xian 710071, Shaanxi, Peoples R China
来源
Wavelet Analysis and Active Media Technology Vols 1-3 | 2005年
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We present a numerical method for measuring the wavelet compressibility of dense linear systems arising from numerical solution of integral equations. The measurement is based on the numerical relation between sparse degree (SD) and relative error (RE), represented by the so-called 'SD-RE curve', which may reflect the properties of both the dense linear system and the wavelet filter under consideration. Near-optimal configuration of wavelet filters and transform levels is hopefully achievable for a given discretized integral equation due to the comparability of the SD-RE curves. Useful trends for very big systems can then be obtained by extrapolation. Computational examples are provided to illustrate the method.
引用
收藏
页码:1138 / 1143
页数:6
相关论文
共 50 条