The Theory Behind and the User's Manual for the Generalized Least Squares Fit (GELSF) Digital Computer Program,

Abstract

In a series of n observations or measurements, y sub i, each of which has k known associated variables (x sub 1,i), (x sub 2,i),...(x sub k,i), it is often possible to construct a linear statistical model to describe the system: y = b sub o + (b sub 1)(x sub 1)+...+(b sub k)(x sub k) + e. Here the (b sub i)'s are fixed but unknown quantities and e is additive random noise. Assuming independent observations with equal variances, the least squares estimates of the coefficients (b sub i)'s are found. If e has a normal (Gaussian) distribution, the least squares estimators are also the maximum likelihood estimators and several well known statistical tests for various hypotheses are applicable. The Generalized Least Squares Fit digital computer program provides the least squares estimates of the unknown coefficients. It also gives the covariance and correlation matrices of the estimates. (Author)

Document Details

Document Type
Technical Report
Publication Date
Oct 23, 1968
Accession Number
AD0846245

Entities

People

  • Nancy R. Rich

Tags

DTIC Thesaurus Topics

  • Additives (Chemicals)
  • Coefficients
  • Computer Programs
  • Computers
  • Covariance
  • Data Science
  • Digital Computers
  • Estimators
  • Hypotheses
  • Information Science
  • Mathematics
  • Measurement
  • Observation
  • Statistical Algorithms
  • Statistical Analysis
  • Statistical Tests

Fields of Study

  • Mathematics

Readers

  • Analytical Mechanics
  • Computer Science.
  • Statistical inference.