CLASSIFICATION AND DISCRIMINATION IN THE ANALYSIS OF CREDIT RISKS: I.
Abstract
Clustering and classification techniques are introduced for the categorization of credit risks. A metric distance for qualitative characteristics is defined. Some shortcomings of clustering by maximum correlation are discussed. Connections between cluster analysis and algorithms of graph theory and of mathematical programming will be considered in a forthcoming supplement to this report. A second report will deal with discrimination of risk categories. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1968
- Accession Number
- AD0666809
Entities
People
- Giandomenico Majone
Organizations
- Carnegie Mellon University