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

Statistical tests for small-sample c...

Guo, Xu.

FindBook

Google Book

Amazon

博客來

Statistical tests for small-sample correlated data with applications to DNA microarray data.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Statistical tests for small-sample correlated data with applications to DNA microarray data./
作者:	Guo, Xu.
面頁冊數:	107 p.
附註:	Source: Dissertation Abstracts International, Volume: 65-03, Section: B, page: 1389.
Contained By:	Dissertation Abstracts International65-03B.
標題:	Statistics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3124746

Statistical tests for small-sample correlated data with applications to DNA microarray data.
Guo, Xu.

Statistical tests for small-sample correlated data with applications to DNA microarray data. - 107 p.

Source: Dissertation Abstracts International, Volume: 65-03, Section: B, page: 1389.

Thesis (Ph.D.)--University of Minnesota, 2004.

Correlated data often arise from repeated measurements on the same subject or clustered sampling. The number of independent subjects or clusters can be small in many situations, such as group-randomized clinical trials, time-course microarray experiments and etc. Statistical inferential techniques based on large-sample approximations may not be appropriate to apply to small-sample correlated data. So, it has become a very active research area to develop statistical techniques for small-sample correlated data to make valid statistical inferences.Subjects--Topical Terms:

517247
Statistics.

Statistical tests for small-sample correlated data with applications to DNA microarray data.
LDR:03394nmm 2200313 4500 001 1862312
005 20041215100250.5
008 130614s2004 eng d
035 $a (UnM)AAI3124746
035 $a AAI3124746
040 $a UnM $c UnM
100 1 $a Guo, Xu. $3 1949874
245 1 0 $a Statistical tests for small-sample correlated data with applications to DNA microarray data.
300 $a 107 p.
500 $a Source: Dissertation Abstracts International, Volume: 65-03, Section: B, page: 1389.
500 $a Advisers: John Connett; Wei Pan.
502 $a Thesis (Ph.D.)--University of Minnesota, 2004.
520 $a Correlated data often arise from repeated measurements on the same subject or clustered sampling. The number of independent subjects or clusters can be small in many situations, such as group-randomized clinical trials, time-course microarray experiments and etc. Statistical inferential techniques based on large-sample approximations may not be appropriate to apply to small-sample correlated data. So, it has become a very active research area to develop statistical techniques for small-sample correlated data to make valid statistical inferences.
520 $a Generalized estimating equations (GEE) has been widely used to model correlated data marginally. However, hypothesis testing based on large-sample approximations does not work well for small-sample situations. First, the small-sample performance of two robust tests in GEE are studied through simulations. Further, the robust tests are modified to achieve better performance for small-sample correlated data.
520 $a Second, GEE is applied to model a special case of small-sample correlated data, longitudinal gene expression data from DNA microarray technologies. A robust statistic is constructed to detect temporal differential gene expression. Statistical significance is defined by using a class of non-parametric methods including Significance Analysis of Microarrays (SAM) method, mixture model method (MMM) and etc. The proposed method is applied to a case study of osteoblast differentiation to identify important genes involved in the osteoblast lineage-specific differentiation pathway.
520 $a The class of non-parametric methods have some potential drawbacks since they all depend on permutations to estimate the null distribution (or characteristic) of the test statistic. However, the standard permutational distribution (or characteristic) may not estimate the null distribution (or characteristic) of the test statistic well, leading to possibly too conservative inferences. Finally, a weighted permutation method is proposed to overcome these drawbacks. Posterior probabilities of having no differential expression from Empirical Bayes (EB) method are used as weights for different genes to obtain better estimates of the null distribution (or characteristic) of the test statistic. As a result, the performance of the class of non-parametric methods are improved with respect to the power to detect differential gene expression and the estimation of false discovery rate (FDR).
590 $a School code: 0130.
650 4 $a Statistics. $3 517247
650 4 $a Biology, Biostatistics. $3 1018416
690 $a 0463
690 $a 0308
710 2 0 $a University of Minnesota. $3 676231
773 0 $t Dissertation Abstracts International $g 65-03B.
790 1 0 $a Connett, John, $e advisor
790 1 0 $a Pan, Wei, $e advisor
790 $a 0130
791 $a Ph.D.
792 $a 2004
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3124746