General Similarity Measures of Location Models,
Abstract
The location models, which can be used in discriminant problems when the data contain both categorical and continuous variables, requires separate continuous variables means to be fitted for each possible pattern of categorical responses. Several forms of similarity measure are reviewed. The problem of estimating similarity when the continuous variables of location models are multivariate normal distributions with equal covariance matrices across the discrete states has previously been studied. In this work, the assumption of equal covariance matrices is relaxed. The explicit form of general similarity measure between two location models is derived assuming general multivariate normal distributions. Estimation of parameters in this similarity measure is discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1992
- Accession Number
- ADP007108
Entities
People
- Ruey-pyng Lu
Organizations
- North Dakota State University