Quantization maximization for learning multiple covariance matrices

碩士 === 國立東華大學 === 應用數學系 === 100 === This work proposes a novel quantization maximization approach for learning mutliple covariance matrices subject to multi-dimensional training data. The number of monitored sources is assumed more than that of sensors. And their statistically dependent relations ar...

Full description

Bibliographic Details
Main Authors:	Shen-Hung Lin, 林紳宏
Other Authors:	Jiann-Ming Wu
Format:	Others
Language:	en_US
Published:	2012
Online Access:	http://ndltd.ncl.edu.tw/handle/36265701825998577510

id	ndltd-TW-100NDHU5507004
record_format	oai_dc
spelling	ndltd-TW-100NDHU55070042015-10-14T04:07:03Z http://ndltd.ncl.edu.tw/handle/36265701825998577510 Quantization maximization for learning multiple covariance matrices 量子化與最大化非監督式學習法解共變異矩陣分析問題 Shen-Hung Lin 林紳宏碩士國立東華大學應用數學系 100 This work proposes a novel quantization maximization approach for learning mutliple covariance matrices subject to multi-dimensional training data. The number of monitored sources is assumed more than that of sensors. And their statistically dependent relations are characterized by multiple covariance matrices. The proposed QM approach iteratively executes quantization and maximization steps. Its convergence is proved and its applicability for spike sorting and underdetermined independent component analysis is extensively explored. Jiann-Ming Wu 吳建銘 2012 學位論文 ; thesis 45 en_US
collection	NDLTD
language	en_US
format	Others
sources	NDLTD
description	碩士 === 國立東華大學 === 應用數學系 === 100 === This work proposes a novel quantization maximization approach for learning mutliple covariance matrices subject to multi-dimensional training data. The number of monitored sources is assumed more than that of sensors. And their statistically dependent relations are characterized by multiple covariance matrices. The proposed QM approach iteratively executes quantization and maximization steps. Its convergence is proved and its applicability for spike sorting and underdetermined independent component analysis is extensively explored.
author2	Jiann-Ming Wu
author_facet	Jiann-Ming Wu Shen-Hung Lin 林紳宏
author	Shen-Hung Lin 林紳宏
spellingShingle	Shen-Hung Lin 林紳宏 Quantization maximization for learning multiple covariance matrices
author_sort	Shen-Hung Lin
title	Quantization maximization for learning multiple covariance matrices
title_short	Quantization maximization for learning multiple covariance matrices
title_full	Quantization maximization for learning multiple covariance matrices
title_fullStr	Quantization maximization for learning multiple covariance matrices
title_full_unstemmed	Quantization maximization for learning multiple covariance matrices
title_sort	quantization maximization for learning multiple covariance matrices
publishDate	2012
url	http://ndltd.ncl.edu.tw/handle/36265701825998577510
work_keys_str_mv	AT shenhunglin quantizationmaximizationforlearningmultiplecovariancematrices AT línshēnhóng quantizationmaximizationforlearningmultiplecovariancematrices AT shenhunglin liàngzihuàyǔzuìdàhuàfēijiāndūshìxuéxífǎjiěgòngbiànyìjǔzhènfēnxīwèntí AT línshēnhóng liàngzihuàyǔzuìdàhuàfēijiāndūshìxuéxífǎjiěgòngbiànyìjǔzhènfēnxīwèntí
_version_	1718090184941109248

Quantization maximization for learning multiple covariance matrices

Similar Items