東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Selected Topics in Statistical Compu...

Zhang, Yiwen.

FindBook

Google Book

Amazon

博客來

Selected Topics in Statistical Computing.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Selected Topics in Statistical Computing./
作者:	Zhang, Yiwen.
面頁冊數:	214 p.
附註:	Source: Dissertation Abstracts International, Volume: 76-04(E), Section: B.
Contained By:	Dissertation Abstracts International76-04B(E).
標題:	Statistics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3647716
ISBN:	9781321414660

Selected Topics in Statistical Computing.
Zhang, Yiwen.

Selected Topics in Statistical Computing. - 214 p.

Source: Dissertation Abstracts International, Volume: 76-04(E), Section: B.

Thesis (Ph.D.)--North Carolina State University, 2014.

This item must not be sold to any third party vendors.

As the connection between statistical models and the real data, optimization methods draw attentions from both academia and industries. Since advances in technology enable easier data collection, complex models are in demand to cope with data that have complex structures; efficient optimization methods are needed to perform analysis on large and ever-growing volumes of data. In this dissertation, we investigate the optimization methods that target at these two issues. The first half of the thesis focuses on developing and implementing optimization algorithms for multivariate generalized linear models (MGLM), while the second half addresses the problems with large data sets.

ISBN: 9781321414660Subjects--Topical Terms:

517247
Statistics.

Selected Topics in Statistical Computing.
LDR:03267nmm a2200313 4500 001 2061337
005 20150929074206.5
008 170521s2014 ||||||||||||||||| ||eng d
020 $a 9781321414660
035 $a (MiAaPQ)AAI3647716
035 $a AAI3647716
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zhang, Yiwen. $3 1032350
245 1 0 $a Selected Topics in Statistical Computing.
300 $a 214 p.
500 $a Source: Dissertation Abstracts International, Volume: 76-04(E), Section: B.
500 $a Adviser: Hua Zhou.
502 $a Thesis (Ph.D.)--North Carolina State University, 2014.
506 $a This item must not be sold to any third party vendors.
520 $a As the connection between statistical models and the real data, optimization methods draw attentions from both academia and industries. Since advances in technology enable easier data collection, complex models are in demand to cope with data that have complex structures; efficient optimization methods are needed to perform analysis on large and ever-growing volumes of data. In this dissertation, we investigate the optimization methods that target at these two issues. The first half of the thesis focuses on developing and implementing optimization algorithms for multivariate generalized linear models (MGLM), while the second half addresses the problems with large data sets.
520 $a Data with multivariate count responses frequently occur in modern applications. The commonly used multinomial-logit model is limiting due to its restrictive mean-variance structure. For instance, analysis of count data from the recent RNA-seq technology by the multinomial-logit model leads to serious errors in hypothesis testing. The ubiquity of over-dispersion and complex correlation structures among multivariate counts call for more flexible regression models. In this dissertation, we study some generalized linear models with multivariate count responses that incorporate various correlation structures among the components of the response vector. Current literature lacks treatment of these models, partly due to the fact that they do not belong to the natural exponential family. We derive stable optimization algorithms under the minorization-maximization (MM) principle. Parameter estimation, testing, and variable selection for these models are treated in a unifying framework. The regression models are compared on both synthetic and real RNA-seq data.
520 $a Traditional iterative optimization algorithms require the whole data set to be available in memory for each iteration. When analyzing large data sets, this optimization scheme makes the computation sensitive to the memory limit. We investigate the online optimization algorithms that provide elegant solutions to the analysis on large data sets. Instead of keeping the whole data set in memory, online algorithms only keep a small batch of data points in memory at each iteration, and process every data point once. We focus on adapting the MM algorithm to the online scheme.
590 $a School code: 0155.
650 4 $a Statistics. $3 517247
650 4 $a Computer science. $3 523869
690 $a 0463
690 $a 0984
710 2 $a North Carolina State University. $b Statistics. $3 2098676
773 0 $t Dissertation Abstracts International $g 76-04B(E).
790 $a 0155
791 $a Ph.D.
792 $a 2014
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3647716