2005 Fiscal Year Final Research Report Summary
Multivariate extreme value theory : Models and estimation
Project/Area Number |
16500174
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Single-year Grants |
Section | 一般 |
Research Field |
Statistical science
|
Research Institution | Kobe University |
Principal Investigator |
TAKAHASHI Rinya Kobe University, Faculty of Maritime Sciences, Professor, 海事科学部, 教授 (80030047)
|
Co-Investigator(Kenkyū-buntansha) |
SIBUYA Masaaki Takachiho University, Faculty of Commercial Sciences, Professor, 商学部, 教授 (20146723)
|
Project Period (FY) |
2004 – 2005
|
Keywords | extreme value theory / multivariate extreme value distribution / statistical model / data analysis / dependence measure / 従属性 / 上位r個 |
Research Abstract |
1.In multivariate extreme value data analysis, the methods of using the upper r largest data and the power transformation of the marginal data were studied. We showed that these methods were efficient for estimating the marginal distributions of the multivariate extreme value distributions. 2.Dependence among components of the multivariate extreme value distributions was characterized by its spectrum measure or by its exponent measure. A dependence order based on stable tail dependence function was introduced. The implication of the order in the risk theory was discussed.
|
Research Products
(6 results)
-
-
-
[Journal Article] The shapes of the probability density, hazard, and reverse hazard functions.2005
Author(s)
Sibuya, M
-
Journal Title
Advance on Models, Characterizations and Applications.(Balakrishnan, N., Bairamov, I.G.and Gebizlioglu, O.L.(Eds))(Chapman, Hall, Boca Raton, FL)
Pages: 1-18
Description
「研究成果報告書概要(欧文)」より
-
-
-
[Book] The shapes of the probability density, hazard, and reverse hazard functions. Balakrishnan, N., Bairamov, I. G. and Gebizlioglu, O. L. (Eds), Advance on Models, Characterizations and Applications.2005
Author(s)
Sibuya, M.
Total Pages
1-19.
Publisher
Chapman Hall, Boca Raton
Description
「研究成果報告書概要(和文)」より