Discrimination, Allocatory and Separatory, Linear Aspects

Abstract

This paper delineates two distinct purposes of discriminatory analyses-- allocation and separation--and examines the linear aspects involved. After a review of multivariate normality and a discussion of the extent to which linearity is optimal, suggestions are made as to the actual use of linear discriminants. Then, dropping distributional assumptions, we focus on the separation of populations via linear functions. We also discuss the application of sample reuse procedures to linear discriminants.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 09, 1976
Accession Number
ADA030698

Entities

People

  • Seymour Geisser

Organizations

  • Stanford University

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Bayesian Networks
  • Classification
  • Computational Science
  • Covariance
  • Data Mining
  • Data Science
  • Discriminant Analysis
  • Information Science
  • Military Research
  • New York
  • Probability
  • Social Psychology
  • Statistical Algorithms
  • Statistical Analysis
  • Statistical Inference
  • Statistics

Readers

  • Business Analytics
  • Regression Analysis.