Methodology and Statistics
International Conference, 24 - 27 September 2000
Hotel Bor and Castle Hrib, Preddvor, Slovenia
Joint Correspondence Analysis
Michael Greenacre
University Pompeu Fabra, Barcelona, Catalonia, Spain
Joint correspondence analysis (JCA) is an adaptation of
multiple correspondence analysis to fit only the off-diagonal
blocks of the Burt matrix, assuming the diagonal blocks are
missing. In this talk we review the properties of JCA and
show how it improves the graphical display of the set of
two-way associations of a set of categorical variables. We
also discuss the dimensionality of the solution, measures of
fit and the role of the numerical contributions in the
interpretation of the solution.
Key words: correspondence analysis, graphical display
See also:
The Millennium Song
|