Factor analysis as a basis for regression procedures with soil parameters

The paper introduces a new procedure for using factor analysis as the basis of the study of regressive relations between individual variables. Factor analysis is here not considered as an alternative to regression analysis, as is the usual case, but rather as a supplement to it. The data to be investigated, taken from a population, are first subjected to factor analysis. This provides the actual factor pattern as well as a matrix with the linear correlation coefficients between the variables. In contrast to the correlation matrix, the factor pattern allows the recognition of all relations existing in the body of data. This includes the recognition of possible dependencies from values which had not been introduced up to that point. Factor analysis allows the recognition and interpretation of structures in the body of data which would not have become evident by the use of the correlation matrix alone (a).

Language

  • English

Media Info

  • Pagination: 59-70
  • Serial:
    • Volume: 1

Subject/Index Terms

Filing Info

  • Accession Number: 01438049
  • Record Type: Publication
  • Source Agency: ARRB
  • ISBN: 0909796084
  • Files: ATRI
  • Created Date: Aug 24 2012 9:20PM