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

Goodness-of-fit tests for proportion...

Luo, Junxiang.

Linked to FindBook

Google Book

Amazon

博客來

Goodness-of-fit tests for proportional odds model with GEE for ordinal categorical responses and estimating sampling frequency in pollen exposure assessment over time.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Goodness-of-fit tests for proportional odds model with GEE for ordinal categorical responses and estimating sampling frequency in pollen exposure assessment over time./
Author:	Luo, Junxiang.
Description:	146 p.
Notes:	Source: Dissertation Abstracts International, Volume: 67-09, Section: B, page: 4792.
Contained By:	Dissertation Abstracts International67-09B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3231106
ISBN:	9780542871962

Goodness-of-fit tests for proportional odds model with GEE for ordinal categorical responses and estimating sampling frequency in pollen exposure assessment over time.
Luo, Junxiang.

Goodness-of-fit tests for proportional odds model with GEE for ordinal categorical responses and estimating sampling frequency in pollen exposure assessment over time. - 146 p.

Source: Dissertation Abstracts International, Volume: 67-09, Section: B, page: 4792.

Thesis (Ph.D.)--University of Cincinnati, 2006.

Chapter I. In 1994, generalized estimating equations approach (GEE) was extended by Lipsitz, Kim and Zhao to fit the proportional odds model for analyzing repeated (or clustered) ordinal categorical data. However, few methods exist to assess the goodness of fit (GOF) of the fitted models. Four U statistics were proposed in this study to do the GOF test and their performances were evaluated with respect to type I error rates and powers for detecting various model departures by simulation studies and an example illustration. We also compared the proposed U statistics with a Kappa-like classification statistic. Simulation studies show that the proposed U statistics have no obvious difference with each other and perform better than Kappa like classification statistic. All of U statistics provide correct type I error rates very close to the nominal alpha levels and appropriate powers to detect the omission of a quadratic term or an interaction, or the violation of proportional odds assumption. Chapter II. A time series model was fitted to the pollen concentration data collected at Greater Cincinnati area for Cincinnati Childhood Allergy and Air Pollution Study (CCAAPS). A traditional time series analysis and temporal variogram approach were applied to the regularly spaced databases (collected in 2003) and irregularly spaced ones (collected in 2002), respectively. The aim was to evaluate the effect of the sampling frequency on the sampling precision in terms of inverse of standard error of the overall level of mean value across time. The presence of high autocorrelation in the data was confirmed and indicated some degree of temporal redundancy in the pollen concentration data. Therefore, it was suggested that sampling frequency could be reduced from once a day to once every several days without a major loss of sampling precision of the grand mean over time. Considering the trade-offs between sampling frequency and the possibility of sampling bias increasing with larger sampling interval, we recommend that the sampling interval should take values from 3 to 5 days for the pollen monitoring program, if the goal is to track the long-term average.

ISBN: 9780542871962Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Goodness-of-fit tests for proportional odds model with GEE for ordinal categorical responses and estimating sampling frequency in pollen exposure assessment over time.
LDR:03151nam 2200277 4500 001 1394036
005 20110419112703.5
008 130515s2006 ||||||||||||||||| ||eng d
020 $a 9780542871962
035 $a (UMI)AAI3231106
035 $a AAI3231106
040 $a UMI $c UMI
100 1 $a Luo, Junxiang. $3 1672620
245 1 0 $a Goodness-of-fit tests for proportional odds model with GEE for ordinal categorical responses and estimating sampling frequency in pollen exposure assessment over time.
300 $a 146 p.
500 $a Source: Dissertation Abstracts International, Volume: 67-09, Section: B, page: 4792.
500 $a Adviser: Rakesh Shukla.
502 $a Thesis (Ph.D.)--University of Cincinnati, 2006.
520 $a Chapter I. In 1994, generalized estimating equations approach (GEE) was extended by Lipsitz, Kim and Zhao to fit the proportional odds model for analyzing repeated (or clustered) ordinal categorical data. However, few methods exist to assess the goodness of fit (GOF) of the fitted models. Four U statistics were proposed in this study to do the GOF test and their performances were evaluated with respect to type I error rates and powers for detecting various model departures by simulation studies and an example illustration. We also compared the proposed U statistics with a Kappa-like classification statistic. Simulation studies show that the proposed U statistics have no obvious difference with each other and perform better than Kappa like classification statistic. All of U statistics provide correct type I error rates very close to the nominal alpha levels and appropriate powers to detect the omission of a quadratic term or an interaction, or the violation of proportional odds assumption. Chapter II. A time series model was fitted to the pollen concentration data collected at Greater Cincinnati area for Cincinnati Childhood Allergy and Air Pollution Study (CCAAPS). A traditional time series analysis and temporal variogram approach were applied to the regularly spaced databases (collected in 2003) and irregularly spaced ones (collected in 2002), respectively. The aim was to evaluate the effect of the sampling frequency on the sampling precision in terms of inverse of standard error of the overall level of mean value across time. The presence of high autocorrelation in the data was confirmed and indicated some degree of temporal redundancy in the pollen concentration data. Therefore, it was suggested that sampling frequency could be reduced from once a day to once every several days without a major loss of sampling precision of the grand mean over time. Considering the trade-offs between sampling frequency and the possibility of sampling bias increasing with larger sampling interval, we recommend that the sampling interval should take values from 3 to 5 days for the pollen monitoring program, if the goal is to track the long-term average.
590 $a School code: 0045.
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Statistics. $3 517247
690 $a 0308
690 $a 0463
710 2 $a University of Cincinnati. $3 960309
773 0 $t Dissertation Abstracts International $g 67-09B.
790 1 0 $a Shukla, Rakesh, $e advisor
790 $a 0045
791 $a Ph.D.
792 $a 2006
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3231106