Message-ID: <E38F7A60-06E0-4DCF-BB28-5F3F6DA0B3B8@math.ucsd.edu>
Date: 2009-05-15T23:18:11Z
From: eariasca
Subject: problem with mle.stepwise
Hello,
I am getting some unexpected results when using mle.stepwise:
# BEGIN
require(wle)
x1 = runif(100)
x2 = runif(100)
y = 3 + 2*x1 + rnorm(100)
# so x1 is in the model, but not x2
mle.stepwise(y ~ x1 + x2, type = "Backward")
# Backward selection gets it right, the other two return non-sense
mle.stepwise(y ~ x1 + x2)
mle.stepwise(y ~ x1 + x2, type = "Stepwise")
# If the variables are entered in a different order, the result changes
mle.stepwise(y ~ x2 + x1)
mle.stepwise(y ~ x2 + x1, type = "Stepwise")
# END
I am running R version 2.6.0 Patched (2007-11-09 r43408) on Mac OS X
10.4.11 and the latest version of wle 0.9-3.
Any idea? Help much appreciated. I will introduce these functions
in my applied statistics class on Monday!
Thanks,
Ery