Clustering Navy Rating by Loss Behavior
Abstract
The enlisted Navy Ratings were clustered by their historical loss behavior, using a hierarchical clustering technique. The immediate application of this clustering technique was to investigate pooling of loss data to improve loss estimation. No significant improvement in loss estimation was found by clustering. Examples of other potential uses for this clustering technique include isolation of groups of ratings to which a common policy regarding loss, reenlistment, etc., may apply.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1975
- Accession Number
- ADA522403
Entities
People
- P. R. Milch
- R. W. Butterworth
Organizations
- Naval Postgraduate School