東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Using simulation to study behaviors ...

Xie, Jieru.

Linked to FindBook

Google Book

Amazon

博客來

Using simulation to study behaviors of parameter estimates in non-linear regression models.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Using simulation to study behaviors of parameter estimates in non-linear regression models./
Author:	Xie, Jieru.
Description:	156 p.
Notes:	Adviser: Linda Jane Goldsmith.
Contained By:	Masters Abstracts International45-01.
Subject:	Biology, Bioinformatics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=1437391
ISBN:	9780542836305

Using simulation to study behaviors of parameter estimates in non-linear regression models.
Xie, Jieru.

Using simulation to study behaviors of parameter estimates in non-linear regression models. - 156 p.

Adviser: Linda Jane Goldsmith.

Thesis (M.S.P.H.)--University of Louisville, 2006.

Non-linear regression models can be used to describe data relationships in more complicated forms than the familiar linear model. In non-linear regression, the least square (LS) estimates of the parameters q&d4; do not have a definite form and they need to be determined by iterative methods. Since the parameter estimates q&d4; in a non-linear model are not a linear combination of observed response Y, the estimates q&d4; do not follow a normal distribution and the properties of q&d4; are unknown. In this case, it's dangerous to use the output from SAS or R to predict the confidence interval for q . This is because the standard error from PROC NLIN (in SAS) and nls (in R) is only appropriate when the parameter estimator is close to Gaussian. In order to get a goodness-of-fit regression model, it is important to study how far the parameter estimators are from having the linear property.

ISBN: 9780542836305Subjects--Topical Terms:

1018415
Biology, Bioinformatics.

Using simulation to study behaviors of parameter estimates in non-linear regression models.
LDR:02509nam 2200301 a 45 001 974410
005 20110929
008 110929s2006 eng d
020 $a 9780542836305
035 $a (UMI)AAI1437391
035 $a AAI1437391
040 $a UMI $c UMI
100 1 $a Xie, Jieru. $3 1298343
245 1 0 $a Using simulation to study behaviors of parameter estimates in non-linear regression models.
300 $a 156 p.
500 $a Adviser: Linda Jane Goldsmith.
500 $a Source: Masters Abstracts International, Volume: 45-01, page: 0206.
502 $a Thesis (M.S.P.H.)--University of Louisville, 2006.
520 $a Non-linear regression models can be used to describe data relationships in more complicated forms than the familiar linear model. In non-linear regression, the least square (LS) estimates of the parameters q&d4; do not have a definite form and they need to be determined by iterative methods. Since the parameter estimates q&d4; in a non-linear model are not a linear combination of observed response Y, the estimates q&d4; do not follow a normal distribution and the properties of q&d4; are unknown. In this case, it's dangerous to use the output from SAS or R to predict the confidence interval for q . This is because the standard error from PROC NLIN (in SAS) and nls (in R) is only appropriate when the parameter estimator is close to Gaussian. In order to get a goodness-of-fit regression model, it is important to study how far the parameter estimators are from having the linear property.
520 $a Complicated methods have been developed to study various measures of non-linear behavior of parameter estimators, including bias and skewness estimation and as well as measures of intrinsic and parameter-effects curvature. In this thesis, the author used Monte Carlo simulation to study the distribution of parameter estimators in two non-linear models which are used often in biological and clinical investigation: the Michaelis-Menten model and the Logistic model. Using simulation, the author also investigated how the data collecting regions affect the behaviors of parameter estimates such as the accuracy, distribution, converge rate and the nonlinearity curvature.
590 $a School code: 0110.
650 4 $a Biology, Bioinformatics. $3 1018415
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Health Sciences, Public Health. $3 1017659
690 $a 0308
690 $a 0573
690 $a 0715
710 2 0 $a University of Louisville. $3 1017614
773 0 $t Masters Abstracts International $g 45-01.
790 $a 0110
790 1 0 $a Goldsmith, Linda Jane, $e advisor
791 $a M.S.P.H.
792 $a 2006
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=1437391