서울대학교 중앙도서관
S-Space 소개
My S-Space
로그인이 필요합니다.
Login

S-Space

Publications

Detailed Information

Typhoon Track Prediction using Bayesian Statistics Methods

DC Field Value Language
dc.contributor.advisor이재용-
dc.contributor.author한민규-
dc.date.accessioned2017-07-14T00:31:30Z-
dc.date.available2017-07-14T00:31:30Z-
dc.date.issued2015-02-
dc.identifier.other000000025552-
dc.identifier.urihttps://hdl.handle.net/10371/121152-
dc.description학위논문 (박사)-- 서울대학교 대학원 : 통계학과, 2015. 2. 이재용.-
dc.description.abstractIn this dissertation we predict the track of typhoons using Bayesian principal component regression model and Bayesian hierarchical regression model based on the wind field data. The data is obtained at each time point and we applied the Bayesian principal component regression model to conduct the track prediction based on the time point. Moreover, we applied the Bayesian hierarchical regression model to conduct the track prediction based on the each typhoon. We show that the results in this dissertation are close to the result of Japanese Meteorological Agency, it is possible to predict the track of typhoon very accurately using only statistical model.-
dc.description.tableofcontentsAbstract i
List of Figures v
List of Tables viii
1 Introduction 1
1.1 Motivation.............................. 1
1.2 Previousworks ........................... 5
2 Bayesian Principal Component Regression Model 9
2.1 Bayesian high-dimensional regression model . . . . . . . . . . . 9
2.1.1 Covariateselectionprior .................. 11
2.1.2 EmpiricalBayesmethod .................. 11
2.2 Posteriorcomputation ....................... 14
2.2.1 BMAandMPMestimators ................ 16
2.2.2 Prediction.......................... 18
2.3 Results................................ 18
2.3.1 Trackprediction ...................... 18
2.4 Concludingremarks......................... 25
3 Bayesian Hierarchical Regression Model 27
3.1 Introduction............................. 27
3.2 Model ................................ 28
3.2.1 Variableselectionprior................... 29
3.3 Posteriorcomputation ....................... 30
3.3.1 Samplingscheme ...................... 31
3.4 Results................................ 32
3.4.1 Datadescription ...................... 32
3.4.2 Fullstepmethod ...................... 33
3.4.3 Plug-inmethod ....................... 38
3.5 ConcludingRemarks ........................ 47
4 Pitman-Yor Process Mixture Model 48
4.1 Model ................................ 48
4.1.1 Regressionpart ....................... 48
4.1.2 Pitman-Yorprocessmixturemodel . . . . . . . . . . . . 49
4.2 Computation ............................ 50
4.2.1 Theconfigurationparameterz............... 51
4.2.2 TheparametergandΣ .................. 51
4.2.3 Thecoefficientparameterβ ................ 52
4.2.4 Theparameterm,τandV ................ 52
4.2.5 Theparameterαandν................... 52
4.3 Simulationstudy .......................... 52
4.3.1 Constructingx ....................... 53
4.3.2 Constructingy ....................... 54
4.3.3 Settingonotherparameters ................ 55
4.3.4 Case:α∼G(12,1),τ=1(fixed)andΣ=I24 . . . . . 55
4.3.5 Case:α∼G(10,1),τ=1(fixed)andΣ=10·I24 . . . 57
4.3.6 Case:α∼G(10,1),τ∼G(2,1)andΣ=10·I24 . . . . 59
4.4 Predictionresult .......................... 60
4.5 Clusteringresultsforwindfielddata . . . . . . . . . . . . . . . 67
Bibliography 74
.1 AppendixofChapter3........................ 76
.1.1 CASE1:θ0sarerandom,βisarenormal . . . . . . . . 76
.1.2 CASE 2 : θ0s are random, βis have variable selection prior(normal)........................ 80
.2 AppendixofChapter4........................ 80
Abstract in Korean 86		-
dc.formatapplication/pdf-
dc.format.extent7552225 bytes-
dc.format.mediumapplication/pdf-
dc.language.isoen-
dc.publisher서울대학교 대학원-
dc.subjectBayesian principal component regression model-
dc.subjectBayesian hierarchical regression model-
dc.subjectTyphoon track-
dc.subjectWind field-
dc.subject.ddc519-
dc.titleTyphoon Track Prediction using Bayesian Statistics Methods-
dc.typeThesis-
dc.description.degreeDoctor-
dc.citation.pagesviii, 86-
dc.contributor.affiliation자연과학대학 통계학과-
dc.date.awarded2015-02-
Appears in Collections:
Files in This Item:

Altmetrics

Item View & Download Count

Show Simple Item Record
Find it @ SNU

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share

qrcode
SNS Share