THE FACTOR ANALYSIS AND THEPRINCIPAL COMPONENTS METHODIN RESULTS PROCESSING OFSCIENTIFIC RESEARCH IN PHYSICALCULTURE AND SPORTS

Abstract

Relevance of the research topic.
Factor analysis and the principal components
method belong to mathematical disciplines
designed to describe multidimensional objects,
and therefore are very widely used in
biological research. However, materials that
help the user, including a specialist in the field
of physical culture and sports, to use
adequately and correctly the modern methods
of multifactorial statistical analysis in practice,
are currently insufficient. The purpose of the
study is to describe the methodological
requirements for the use of factor analysis
when conducting research in the field of
physical culture and sports. Methods: analysis
of scientific and methodical literature, factor
analysis. Results. Factor analysis does not
require an a priori division of variables into
dependent and independent, and all variables
in it are considered equal. Within factor
analysis, there is no need to assume
independence (non-correlation) of variables,
as for example in regression or correlation
analyses. On the contrary, it is assumed that
these initial variables are interdependent.
There are no restrictions in factor analysis on
the number of features that are initially taken
into account. And the better factors for factor
analysis become the more of them, the more
informative. A priority aspect of organizing
and performing factor analysis is establishing
the number of factors sufficient to adequately
represent the set of variables. An example of
the factor analysis use in determining the
factor structure of the components of the
biogeometric posture profile of women aged 36–39 years with the Statistica 14.0
application software was provided.
Conclusions. Factor analysis and the principal
components method are ultimately designed to
reduce the dimensionality of the primary
feature space. Factor analysis and the principal
components analysis have a common strategy
of application, but a different technology for
achieving the goal. The main components are
determined on the basis of the variance
criterion, and the general factors are
constructed in such a way as to maximally
explain the existing correlations between the
features.