View : 1029 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.author정윤영-
dc.creator정윤영-
dc.date.accessioned2016-08-25T02:08:57Z-
dc.date.available2016-08-25T02:08:57Z-
dc.date.issued2007-
dc.identifier.otherOAK-000000020697-
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/173191-
dc.identifier.urihttp://dcollection.ewha.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000020697-
dc.description.abstract메타자료와 패널자료에 대한 분석은 통계 테스트의 검정력을 증가시켜 불확실성을 감소시키며, 통계적 추정의 정확성을 향상시키기때문에 서로 다른 연구 결과들을 결합하여 통합된 정보를 이끌어 내는데 널리 사용되어왔다. 본 논문은 베이지안 계층 모형을 이용해 유전자 분야에서 발생하는 메타자료를 분석하는 방법을 제시하고 있으며, 비선형 패널 자료를 위한 MTAR 모형과 오랜 시간 서서히 변화해가는 패널 자료 분석을 통해 얻어진 통합된 정보는 개별 분석으로부터 얻어지는 정보에 비해 보다 명확한 결과를 제시하였다. 본 논문에서 제시되는 방법들은 메타 자료와 패널 자료 분야에서 통계기법을 한단계 발전시키는데 기여할 것으로 기대한다.;Meta data analysis has been widely used to combine results from different studies, since it usually increases powers of statistical tests, resolve uncertainty, and improve estimation accuracy. Similarly, panel data modelling is useful for controlling individual heterogeneity of cross-section units and producing more reliable estimates of model parameters. This work propose a Bayesian hierarchical model for analyzing meta data in genomic field and panel versions of a momentum threshold autoregressive (MTAR) model and a double unit root model for nonlinear panel data and slowly changing long-run panel data, respectively. The integrated information through the proposed meta and panel data analysis deliver clearer message than information from a componentwise analysis. The proposed methods will hopefully contribute to bring panel data or meta data fields a signifcant step forward.-
dc.description.tableofcontentsAbstract v Chapter I. Identifying Differentially Expressed Genes in Meta-Analysis via Bayesian Model-Based Clustering 1 1.1 Introduction 1 1.2 Bayesian hierarchical model and a mixture prior 4 1.3 Posterior inference via MCMC 7 1.4 False discovery rate 9 1.5 Analysis of prostate cancer data 11 a) Posterior inference 12 b) False discovery rate 16 c) Integration-driven discovery 18 1.6 Concluding remarks 21 Chapter II. Bayesian Analysis of Panel Data Using an MTAR Model 25 2.1 Introduction 25 2.2 Model and prior specification 27 2.3 Posterior inference 30 2.4 Model selection 33 a) Selection of differenced variables 34 b) Test for unit root 35 c) Test for asymmetry 36 2.5 An example 37 a) Selection of differenced variables 38 b) Estimation of MTAR model 40 c) Testing for Unit Root and Asymmetry 43 2.6 Concluding remarks 44 Chapter III. Double Unit Root Tests for Cross-Sectionally Dependent Panel Data 48 3.1 Introduction 48 3.2 A Model for Cross-Sectionally Dependent Panel Data 51 3.3 Panel Double Unit Root Tests 52 a) Tests based on projection defactoring 53 b) Test based on subtraction defactoring 57 c) Limiting distributions of the test statistics 58 3.4 A Monte Carlo study 60 a) Experimental design 60 b) Simulation results 64 3.5 An example 67 3.6 Concluding remarks 72 REFERENCES 73 국문초록 80-
dc.formatapplication/pdf-
dc.format.extent4007091 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.title메타 자료와 패널자료에 대한 연구-
dc.typeDoctoral Thesis-
dc.title.translatedThree Studies on Meta and Panel Data: Gene Identification, Panel Dynamic Asymmetry, and Panel Double Unit Root-
dc.creator.othernameJung, Yoon Young-
dc.format.pagevii, 82 p.-
dc.identifier.thesisdegreeDoctor-
dc.identifier.major대학원 통계학과-
dc.date.awarded2007. 2-
Appears in Collections:
일반대학원 > 통계학과 > Theses_Ph.D
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE