对称重尾源盲信号分离的超高效率

信号处理 Pub Date : 2001-08-06 DOI:10.1109/SSP.2001.955226

Yoav Shereshevski, A. Yeredor, H. Messer

{"title":"对称重尾源盲信号分离的超高效率","authors":"Yoav Shereshevski, A. Yeredor, H. Messer","doi":"10.1109/SSP.2001.955226","DOIUrl":null,"url":null,"abstract":"This paper addresses the blind source separation (BSS) problem in the context of \"heavy-tailed\", or \"impulsive\" source signals, characterized by the nonexistence of finite second (or higher) order moments. We consider Pham's (1997) quasi-maximum likelihood (QML) approach, a modification of the maximum likelihood (ML) approach, applied using some presumed distributions of the sources. We introduce a related family of suboptimal estimators, termed restricted QML (RQML). A theoretical analysis of the asymptotic performance of RQML is presented. The analysis is used for showing that the variance of the optimal (non-RQML) estimator's error must decrease at a rate faster than 1/T (where T is the number of independent observations). This surprising property, sometimes called super-efficiency, has been observed before (in the BSS context) only for finite-support source distributions. Simulation results illustrate the good agreement with theory.","PeriodicalId":70952,"journal":{"name":"信号处理","volume":"46 1","pages":"78-81"},"PeriodicalIF":0.0000,"publicationDate":"2001-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"9","resultStr":"{\"title\":\"Super-efficiency in blind signal separation of symmetric heavy-tailed sources\",\"authors\":\"Yoav Shereshevski, A. Yeredor, H. Messer\",\"doi\":\"10.1109/SSP.2001.955226\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This paper addresses the blind source separation (BSS) problem in the context of \\\"heavy-tailed\\\", or \\\"impulsive\\\" source signals, characterized by the nonexistence of finite second (or higher) order moments. We consider Pham's (1997) quasi-maximum likelihood (QML) approach, a modification of the maximum likelihood (ML) approach, applied using some presumed distributions of the sources. We introduce a related family of suboptimal estimators, termed restricted QML (RQML). A theoretical analysis of the asymptotic performance of RQML is presented. The analysis is used for showing that the variance of the optimal (non-RQML) estimator's error must decrease at a rate faster than 1/T (where T is the number of independent observations). This surprising property, sometimes called super-efficiency, has been observed before (in the BSS context) only for finite-support source distributions. Simulation results illustrate the good agreement with theory.\",\"PeriodicalId\":70952,\"journal\":{\"name\":\"信号处理\",\"volume\":\"46 1\",\"pages\":\"78-81\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2001-08-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"9\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"信号处理\",\"FirstCategoryId\":\"1093\",\"ListUrlMain\":\"https://doi.org/10.1109/SSP.2001.955226\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"信号处理","FirstCategoryId":"1093","ListUrlMain":"https://doi.org/10.1109/SSP.2001.955226","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 9

摘要

本文研究了“重尾”或“脉冲”源信号中不存在有限秒(或更高)阶矩的盲源分离(BSS)问题。我们考虑了Pham(1997)的准最大似然(QML)方法，这是对最大似然(ML)方法的一种修正，它使用了一些假定的源分布。我们引入了一类相关的次优估计器，称为受限QML (RQML)。对RQML的渐近性能进行了理论分析。该分析用于显示最优(非rqml)估计器误差的方差必须以快于1/T(其中T是独立观测值的数量)的速率减小。这种令人惊讶的特性有时被称为超效率，以前(在BSS上下文中)只在有限支持的源分布中被观察到。仿真结果与理论结果吻合较好。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Super-efficiency in blind signal separation of symmetric heavy-tailed sources

This paper addresses the blind source separation (BSS) problem in the context of "heavy-tailed", or "impulsive" source signals, characterized by the nonexistence of finite second (or higher) order moments. We consider Pham's (1997) quasi-maximum likelihood (QML) approach, a modification of the maximum likelihood (ML) approach, applied using some presumed distributions of the sources. We introduce a related family of suboptimal estimators, termed restricted QML (RQML). A theoretical analysis of the asymptotic performance of RQML is presented. The analysis is used for showing that the variance of the optimal (non-RQML) estimator's error must decrease at a rate faster than 1/T (where T is the number of independent observations). This surprising property, sometimes called super-efficiency, has been observed before (in the BSS context) only for finite-support source distributions. Simulation results illustrate the good agreement with theory.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

信号处理

自引率

0.00%

发文量

5812

期刊介绍： Journal of Signal Processing is an academic journal supervised by China Association for Science and Technology and sponsored by China Institute of Electronics. The journal is an academic journal that reflects the latest research results and technological progress in the field of signal processing and related disciplines. It covers academic papers and review articles on new theories, new ideas, and new technologies in the field of signal processing. The journal aims to provide a platform for academic exchanges for scientific researchers and engineering and technical personnel engaged in basic research and applied research in signal processing, thereby promoting the development of information science and technology. At present, the journal has been included in the three major domestic core journal databases "China Science Citation Database (CSCD), China Science and Technology Core Journals (CSTPCD), Chinese Core Journals Overview" and Coaj. It is also included in many foreign databases such as Scopus, CSA, EBSCO host, INSPEC, JST, etc.