Seemingly a bug in lassoglm?
1 view (last 30 days)
Show older comments
Timothy Scharf
on 19 Aug 2014
Answered: Timothy Scharf
on 19 Aug 2014
Hello all,
rng('default') % for reproducibility
X = randn(100,20);
mu = exp(X(:,[5 10 15])*[.4;.2;.3] + 1);
y = poissrnd(mu);
[B FitInfo] = lassoglm(X,y,'poisson','CV',10);
Error using zscore
Too many input arguments.
Error in lassoglm>computeLambdaMax (line 1509)
[X0,~,~] = zscore(X,1);
Error in lassoglm (line 246)
[lambdaMax, nullDev, nullIntercept] =
computeLambdaMax(X, Y, pwts, alpha, standardize, ...
It looks as if the internal call to zscore is using an improper number of arguments
the call
zscore(X,1)
should be
zscore(X, flag, 1)
if I set the 'Standardize' to false, it runs fine
[B FitInfo] = lassoglm(X,y,'poisson','CV',10,'Standardize',false);
I am using student R2013a and this error occurs the same way on all three examples
Thanks if anyone does a sanity check and confirms -
0 Comments
Accepted Answer
Sean de Wolski
on 19 Aug 2014
which -all zscore
More than likely you have your own zscore function shadowing the builtin. Please rename this one and the above should work.
0 Comments
More Answers (1)
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!