Dear list, This little question of mine has to do with simulating multivariate normals. It's been nagging me for sometime and I'd like to hear the list's comment. Splus has this build-in rmvnorm func
I'm not sure how efficient it is, but here is a way of passing a matrix/dataframe. First, you create a list in Splus with as many elements as no. of columns in your dataframe/matrix. Then you read b
Dear list, This should be very simple and I've wasted enough time on it but still get nowhere. I got a dataframe with NAs scattering all over the place. I want to pick out all rows, or similarly all
I'd like to thank Richard Heiberger, Gerald Jean, Sarah Goslee, Bill Dunlap, and lastly Terry Therneau for their solutions. Trick is to use apply(), subs <- apply(is.na(tmp), 1, function(x) !any(x))
ifelse returns one of the two arguments which must be a vector. if the satisfied argument is a matrix then it returns the first column. Horace >>> "Laura Holt" <lauraholt_983@hotmail.com> 03/17/05 9:
A new addition to the literature is C. Bishop (1995) Neural Networks for Pattern Recognition, Clarendon Press, Oxford. The chapter on NN in Hastie, et. al. Elements of Stat. Learning, Springer is als
Patrick, in my S.h, which really just points to 'S_define.h', I found two macros defined in terms of two different functions, 1. #define Calloc(n,t) (t *)S_ok_calloc((n),sizeof(t), S_evaluator) 2. #d
Gareth, I asked about a similar question a while back and the code is expand.grid(rep(list(0:1), n)) Then I was reminded by a very pragmatic Bill Venables that in the real world where anything beyond