Thanks, Simon. I definitely agree, I ran into quite a few errors but were helpful for me to figure what's wrong.
I did switch the 2012 MacBook Pro, and the problem was gone. My actually data is much larger. But, I did copy paste the error, maybe that's why? Mike On Tue, Dec 18, 2012 at 7:27 PM, Simon Urbanek <[email protected]>wrote: > Mike, > > first, as the error tells you, your "y" is not what svm.fs() requires, > i.e. it has to be a vector of -1 and 1, so in your code that would be > > y <- sample(c(-1,1), 20, replace=TRUE) > > Then further penalizedSVM is seriously buggy - the author fails to use > match.arg() on the arguments, so what happens is that all the defaults are > invalid and break. The consequence is that the user has to specify all > defaults explicitly, otherwise the code doesn't work (interestingly the > author specifies everything in the examples - otherwise it would not pass > checks at all). For example this seem to sort of work (still with some > warnings): > > model <- > svm.fs(x,y,"scad","interval",inner.val.method="cv",parms.coding="none",show="final") > > Given those trivial mistakes in the code and the fact that it uses print() > all over the place which is plain wrong, I'd be very careful with relying > on the penalizedSVM package, so use at your own risk. > > Finally, this is not a Mac question, please consider using R-help for such > queries. (And, no, it has nothing to do with memory usage). > > Cheers, > Simon > > > On Dec 18, 2012, at 6:48 PM, C W wrote: > > > Dear list, > > I am trying to run svm.fs() in package penalizedSVM, > > > > x <- matrix(rnorm(100), 20, 5) > > > > y <- sample(c(0,10), 20, replace=TRUE) > > > > model <- svm.fs(x,y) > > > > And I get the following warnings and errors. > > > > rror in array(0, c(n, p)) : 'dim' specifies too large an array > > > > In addition: There were 16 warnings (use warnings() to see them) > > > > There were 50 or more warnings (use warnings() to see the first 50) > > > >> warnings() > > > > Warning messages: > > > > 1: In if (fs.method == "1norm" & inner.val.method == "gacv") stop("gacv > is > > not availible for 1norm SVM. Use k fold cv instead.") : > > > > the condition has length > 1 and only the first element will be used > > > > 2: In if (grid.search == "interval") { ... : > > > > the condition has length > 1 and only the first element will be used > > > > 3: In if ((fs.method %in% c("scad", "1norm")) & is.null(bounds)) { ... : > > > > the condition has length > 1 and only the first element will be used > > > > 4: In if (fs.method %in% c("DrHSVM") & is.null(bounds)) { ... : > > > > > > I am guessing this is because of the limit memory? > > > > Mike > > > > [[alternative HTML version deleted]] > > > > _______________________________________________ > > R-SIG-Mac mailing list > > [email protected] > > https://stat.ethz.ch/mailman/listinfo/r-sig-mac > > > > > > [[alternative HTML version deleted]] _______________________________________________ R-SIG-Mac mailing list [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-mac
