Concordant and Discordant Monotone Correlations and Their Evaluation by Nonlinear Optimization.
Abstract
This paper presents four new statistical measures of monotone relationship derived from the concept of monotone correlation. A nonlinear optimization algorithm is employed to evaluate these new measures, as well as the monotone correlation, for ordinal contingency tables. A computer program to implement the algorithm is developed, and is applied to several insightful examples to provide further understanding of the usefulness of these measures. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1980
- Accession Number
- ADA093816
Entities
People
- Allan R. Sampson
- George Kimeldorf
- Jerrold H. May
Organizations
- University of Pittsburgh