Statistical Inference for the Skewed-Normal and Related Distributions.

Abstract

The skewed normal distribution provides a useful mechanism for investigating the mode and skewness of a random sample of observations. Marginal likelihood/approximate Bayes techniques are proposed for making inferences about the location scale, and skewness parameters of this distribution. The distribution may also be used for the error terms in the linear statistical model, yielding an alternative to the least squares criterion, taking account of possible skewness in the residuals. A skewed logistic distribution is discussed in relation to bioassay, and a skewed multivariate normal distribution is introduced. The ideas are used to analyze a set of medical data relating to fetal, asphyxia, exmploying skewed normal assumptions combined with a discriminant analysis.

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1982
Accession Number
ADA120991

Entities

People

  • L. Broekhoven
  • Tom Leonard

Organizations

  • University of Wisconsin–Madison

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Bioassay
  • Computing-Related Activities
  • Data Science
  • Discriminant Analysis
  • Information Science
  • Interdisciplinary Science
  • Mathematical Analysis
  • Mathematics
  • Normal Distribution
  • Observation
  • Residuals
  • Skewness
  • Statistical Inference
  • Statistical Samples

Fields of Study

  • Mathematics

Readers

  • Regression Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference