项目名称: 函数重构理论及其在生物信息学中的应用
项目编号: No.10801136
项目类型: 青年科学基金项目
立项/批准年度: 2009
项目学科: 无线电电子学、电信技术
项目作者: 冼军
作者单位: 中山大学
项目金额: 17万元
中文摘要: 本项目研究了加权的平移不变函数空间和多生成平移不变函数空间中的非均匀采样集条件和函数重构算法及其在生物信息学中的应用。当采样集满足某些条件时,才可能获得稳定的函数重构(恢复),我们试图在最一般的平移不变函数空间中获得此条件。在获得非均匀采样条件的基础上,将给出有效的函数重构算法,此算法将具有一般性和更好的收敛性,而且有显式的收敛阶,以便我们找到控制收敛速度的各种因素。最后我们将研究结果,即函数重构算法应用于生物信息学中缺失DNA 芯片数据的处理。理论上可多做一些试验来增加基因表达数据的准确性和完整性,但一些DNA 芯片数据的试验是非常昂贵的,不能经常重复。对于适用于均匀采样情况的一些方法,在对非均匀采样和缺失数据处理已不再适用。我们的方法克服了这些不足。该项目的研究内容不但是小波分析理论、采样理论以及函数逼近论的实质性扩充,而且在生物信号处理及电子信息工程等领域中均有重要的应用。
中文关键词: 小波分析;细分方程;非均匀采样;函数重构;谱分析
英文摘要: This project is research in weighted shift-invariant and weighted multiply generated shift-invariant spaces, that is condition of non-uniformly sampling set and algorithm of function reconstruction in weighted shift-invariant space and multi-generated shiftinvariant space. Their applications are discussed in bioinformatics too. Our goal is to quantify the conditions under which it is possible to obtain the stable signa reconstruction in general shift-invariant space. Under obtaining the condition of nonuniformly sampling set, we hope to obtain an effective reconstruction algorithm that it show the better convergence feature and explicit convergence rate. We can make use of this advantage to control convergence rate. In the end, we will apply our results to deal with missing DNA chip data with algorithm of function reconstruction. In theory, we can do more experiments to improve the veracity and the integrity of DNA expression data. But the cost is expensive. The classical uniformly sampling methods are NOT fit to non-uniformly sampling set and missing data. Our methods are promising in overcoming the shortcomings. In a word, not only this project is the essential generalization of wavelet theory, sampling theory and approximation theory, but also we believe that there are promising applications in biologic signal processing, electronic engineering and so on.
英文关键词: wavelet analysis;refinement equation;irregularly sampling; function reconstrucion;spectral analysis