ANOVA
- To: mathgroup at yoda.physics.unc.edu
- Subject: ANOVA
- From: David Withoff <withoff>
- Date: Fri, 17 Apr 1992 10:16:20 -0500
Regarding your mathgroup posting about ANOVA in Mathematica, a few years
ago I started on an ANOVA package, and got as far as one-way and balanced
two-way designs (that is, not very far) before I moved on to other things.
I'd be happy to send you the package if you are interested. I had two
problems:
1) The mathematics isn't very difficult, but since I am not a trained
statistician I don't know how to design the user interface in a way
that will make sense to statisticians. If someone could write up
a design specification, I'm sure I could write the package (or arrange
to have someone write the package) in about a month.
2) The design matrices that arise from multi-way problems are often large
sparse matrices, and Mathematica doesn't have any special facilities
for handling such matrices. Based on a few experiments, it seemed
to me that the package would only be useful for relatively small problems.
There is a plan to improve Mathematica in this regard.
Dave Withoff
withoff at wri.com