 |
 |
CU Home > Libraries Home | About | Help |
|
|
Columbia Dissertations and Theses > Doctoral Dissertations
|
|||||||||||||||||||||||||||||||||||||||
| Author(s): | Abayomi, Kobi Ako |
| Title: | Diagnostics for multivariate imputation, copula based Independent Component Analysis and a motivating example. |
| Advisor(s): | de la Pena, Victor |
| Physical Description: | 88 p. |
| Issue Date: | 2008 |
| Description: | Source: Dissertation Abstracts International, Volume: 69-01, Section: B, page: 0395. Adviser: Victor de la Pena. Thesis (Ph.D.)--Columbia University, 2008. |
| Bookmark as: | http://hdl.handle.net/10022/AC:P:18421 |
| Full Text (ProQuest): | /ac/proxit.jsp?url=http://gateway.proquest.com/ope... |
| Abstract: | We consider new methods in imputation diagnosis and component extraction using the Environmental Sustainability Index (ESI) - an aggregation of environmental variables created as a measure of overall progress towards environmental sustainability - as a motivating example. In Chapter 2 we consider three sorts of diagnostics for random imputations; we also consider a recalibration procedure for sequential regression imputations. In Chapter 3 we propose a parametric version of Independent Component Analysis (ICA) via Copulas - families of multivariate distributions that join univariate margins to multivariate distributions. In Chapter 4 we apply a version of the CICA approach to the ESI dataset. |
| Collection(s): | Doctoral Dissertations |
Copyright: All rights reserved.