Q-type factor analysis
From the help page:
'princomp' only handles so-called R-mode PCA, that is feature
extraction of variables. If a data matrix is supplied (possibly
via a formula) it is required that there are at least as many
units as variables. For Q-mode PCA use 'prcomp'.
On Wed, 24 Oct 2007, "Julia Kr?pfl" wrote:
Hi there! I have tried your idea with rotating the matrix and performing a normal PCA, but the problem is, that "princomp" can only perform PCA if there are more rows than columns. When I rotate the matrix, I get my observations put in the columns and my features in the rows (more columns than rows) and therefore get an error message. Any ideas what to do? Thx for your help, I really appreciate it! Julia -------- Original-Nachricht --------
Datum: Fri, 12 Oct 2007 23:38:01 +0300 Von: "Kenn Konstabel" <lebatsnok at gmail.com> An: "Julia Kr?pfl" <jkroepfl at gmx.net> CC: r-help at r-project.org Betreff: Re: [R] Q-type factor analysis
On 10/12/07, "Julia Kr?pfl" <jkroepfl at gmx.net> wrote:
Is there a package in R that does Q-type factor analysis? I know how to do principal component analysis, but haven't found any application of Q-type factor analysis.
Q-mode factor analysis is not a separate "type" of factor analysis but (in old-fashioned psychological slang) analyzing of rows rather than the columns of data matrix. So you can transpose your data (with t() if it's a matrix) and do an "ordinary" factor analysis or PCA. Kenn
Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595