[BioC] Estimate how much a factor contributes to the variation
paulgeeleher at gmail.com
Mon Mar 25 21:50:54 CET 2013
Perhaps and R-squared change test (also known as an F-change test) may
be useful. It will allow you to compare two models, i.e. one with and
one without your co-efficient of interest and tell you if the fit is
improved. Can be implemented using the anova() function in R. I.e.
anova(fullModel, reducedModel) I'm fairly sure is the correct syntax.
On Mon, Mar 25, 2013 at 2:58 PM, Robert Svensen
<svensen.robert at gmail.com> wrote:
> Dear Bioconductors,
> I'm analyzing a microarray data set with three factors; cancer (yes or no,
> the interesting one), age and % tumour cells in sample.
> I'm interested to know how much each factor/coefficent is able to describe
> the overall variation in the data set.
> My googlefu turns up little of interest, could someone suggest a suitable
> package or some key words that could take me further?
> [[alternative HTML version deleted]]
> Bioconductor mailing list
> Bioconductor at r-project.org
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
Dr. Paul Geeleher, PhD (Bioinformatics)
Section of Hematology-Oncology
Department of Medicine
The University of Chicago
900 E. 57th St.,
KCBD, Room 7144
Chicago, IL 60637
More information about the Bioconductor