Imputation of missing values using density estimation
The paper proposes the use of density estimation, based on respondents' data, for imputing values for nonrespondents. Multiple imputations are catered for and the procedure is assessed on the basis of estimation of the underlying density itself. Extension to Rubin's version of multiple imputations is described.
| Year of publication: |
1989
|
|---|---|
| Authors: | Titterington, D. M. ; Sedransk, J. |
| Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 8.1989, 5, p. 411-418
|
| Publisher: |
Elsevier |
| Keywords: | kernel method multiple imputations smoothed bootstrap |
Saved in:
Saved in favorites
Similar items by person
-
Analytical surveys with cluster sympling
Sedransk, J., (1965)
-
A double sampling scheme for analytical surveys
Sedransk, J., (1965)
-
Sampling problems in the estimation of the money supply
Sedransk, J., (1977)
- More ...