A Graph-based Approach to Key Variable Mapping

Abstract

This paper describes the Key Variable Mapping System (KVMS) - a metadata repository, and tools for carrying out disclosure risk assessments. The KVMS repository contains metadata relating to the collection instruments used to populate private databases, and partition graphs for encoding possibilities for record matching. Analyses are conducted to assist in the identification of key variables. The data entry system used to populate the repository is described, along with details of the data structures and algorithms used to query the repository and generate a particular form of analysis. KVMS provides a framework for disclosure risk analyses that depend on a prevailing data environment.

Keywords

• privacy
• disclosure
• data environment
• metadata
• graphs
• algorithms