Re: An introduction to the lasso in Stata

Thanks for this nice description of the Lasso tools on stata. One question I have: why do you need to create a split sample at the beginning using:
splitsample , generate(sample) split(.75 .25) rseed(12345)
while the 10-fold crossvalidiation is going to do the job in a much more systematic way. Can't you just run the program without the if sample==1 ?
Thanks