FACTOR ANALYSIS OF DATA MATRICES. PART III,

Abstract

This is Part III f a series of reports on rationales and techniques of matrix factoring which play an important role in multivariate analysis techniques. The methods discussed are basic structure solutions. They include: Jacobi type, order reduction, solutions from incomplete covariance matrices, factoring data matrices. The most important characteristic of these methods is that the solution involves repeated operations with a one-parameter square orthonormal matrix, as distinguished from repeated matrix by vector multiplication. The solutions are interactive or consist of a series of sucdessive approximations. They are not selfcorrecting for the successive approximations. This means that each cycle of computations accumulates decimal error and passes it on to the next cycle. Two variations of the method which involves repeated operations with a single parameter square orthonormal matrix are discussed. The simultaneous method is an iterative solution involving one-parameter orthonormal transformations, or what is sometimes known as binary rotations; the successive method is that which involves the same type of orthonormal transformations of the original matrix, but the basic diagonals and corresponding orthonrmal vectors are solved for in order of magnitude.

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1964
Accession Number
AD0430593

Entities

People

  • Paul Horst

Organizations

  • University of Washington

Tags

DTIC Thesaurus Topics

  • Computational Processes
  • Computations
  • Computing-Related Activities
  • Covariance
  • Data Analysis
  • Data Mining
  • Data Science
  • Factor Analysis
  • Information Science
  • Interdisciplinary Science
  • Mathematical Analysis
  • Mathematics
  • Multivariate Analysis
  • Rotation

Fields of Study

  • Mathematics

Readers

  • Applied Combinatorial Optimization and Logic Circuit Design.
  • Calculus or Mathematical Analysis
  • Regression Analysis.