Categorical proxy data fusion Seminar
- Date:
- 17 October 2012
- Venue:
- Building 58 Room 1039
For more information regarding this seminar, please email Mrs Jane Revell at j.revell@southampton.ac.uk .
Event details
Methodology seminar
Abstract
Conceptually, in data fusion or statistical matching, one is interested in the joint distribution of two variables which are only observed marginally each on its own. In proxy data fusion, we assume that there exist proxy measures of the target ones in addition to the potential auxiliary variables. For instance, while age, sex and region may be considered as auxiliaries to the target ILO Unemployment Status and highest Level of Education, register-based claimant counts may be used to compile a proxy Unemployment Status, and various examination results from the school system may be used to construct a proxy Level of Education. Both proxy variables have the same classification as the respective target ones, despite the differences in concept. In this talk, I will investigate how data fusion may be improved given such proxy measures.
Speaker information
Professor Li-Chun Zhang ,Professor of Social Statistics