Mail Archive Search: <+from:bert

Results:

References: [ +from:bert_gunter@merck.com: 120 ]

Total 120 documents matching your query.

1. Re: S-PLUS Vs some other softwares (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 2 Mar 2004 08:57:01 -0500; Andy said, lmList will probably do what you want, but as it uses lm(), it may also take a while. If you wish to do it "by hand" yourself, try by() [which is a wrapper for tapply()] and use lsfit ins; /archives/html/s-news/2004-03/msg00023.html (13,494 bytes)
2. Re: scalability (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Fri, 26 Mar 2004 16:26:48 -0500; Get V&R's S-PROGRAMMING book. They discuss these issues in detail (as does S-Plus's documentation). Basically, anything that loops in S rather than in the underlying C or Fortran code will not scale; /archives/html/s-news/2004-03/msg00216.html (10,424 bytes)
3. s (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Fri, 16 Jan 2004 08:39:12 -0500; To all: As always, I appreciate Alan's and Spencer's thoughful comments. I find them always educational and often entertaining (which goes to show, I suppose, how truly nerdy I am). May I just add a; /archives/html/s-news/2004-01/msg00094.html (11,892 bytes)
4. FW: filling holes in its series (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Wed, 03 Dec 2003 09:09:37 -0500; One approach is to use match() and diff(). I'm not exactly sure how time series work in S-Plus 4.5 (I use R now), but you should be able to do something like this: Assume its1 and its2 are vectors in; /archives/html/s-news/2003-12/msg00007.html (11,376 bytes)
5. Re: Help on comparing models (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 9 Dec 2003 08:49:25 -0500; anova() only works for testing nested models. So this turns out to be another one of those innocent, "simple" questons beneath which lies a den of statistical snakes. There is a large literature on; /archives/html/s-news/2003-12/msg00043.html (12,515 bytes)
6. Re: stacking columns (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Wed, 22 Oct 2003 08:12:38 -0400; 6 Oct 2003, Sundar Dor; /archives/html/s-news/2003-10/msg00218.html (10,057 bytes)
7. Re: plackett burman design (score: 1): Author: m>; Date: Wed, 22 Oct 2003 10:51:04 -0400; Statistics for Experi; /archives/html/s-news/2003-10/msg00221.html (10,119 bytes)
8. Re: help with a dataframe (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 13 May 2003 08:39:18 -0400; be; /archives/html/s-news/2003-05/msg00048.html (11,331 bytes)
9. Re: extracting components of output (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 25 Mar 2003 07:58:20 -0500; ory 707-; /archives/html/s-news/2003-03/msg00175.html (9,662 bytes)
10. At the risk of overkill ... (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Thu, 27 Mar 2003 13:21:49 -0500; te("samp; /archives/html/s-news/2003-03/msg00196.html (7,546 bytes)
11. Re: About_Optimization (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 11 Feb 2003 12:22:56 -0500; I don't know about speed, but this kind of example is the sort of thing for which outer() seems ideally suited: x<-colSums(outer(z,y,"<")) Maybe slower but hopefully more transparent than the other p; /archives/html/s-news/2003-02/msg00064.html (8,658 bytes)
12. Re: Multivariate Kernel Densities (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Thu, 16 Jan 2003 08:07:50 -0500; FWIW, my experience is that it is often reasonable to project down to low dimensional (linear, one hopes) manifolds (even 1 dimension often suffices) and visualize the data there. So I don't think ma; /archives/html/s-news/2003-01/msg00072.html (11,920 bytes)
13. Re: moving median in S without looping -- again! (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 21 Jan 2003 15:33:25 -0500; To anyone who can stand one more: I replied privately to the original poster, but there have been so many suboptimal/complicated solutions presented, that I feel I need to post it publicly. First, an; /archives/html/s-news/2003-01/msg00115.html (13,630 bytes)
14. Typos -- Dumb Dumb! (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 21 Jan 2003 16:35:35 -0500; I apologize for my typos. The corrected version of the moving median trick is (I hope!): n<-length(z) z.stack<- c(z[(1:(n-2)],z[2:(n-1)],z[3:n]) indx<-rep(1:(n-2),3) z.ord<-z.stack[order(indx,z.stack; /archives/html/s-news/2003-01/msg00118.html (7,131 bytes)
15. Re: mode of objects (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Wed, 04 Dec 2002 08:01:52 -0500; ... or yet more simply: sapply(objects(),function(x)mode(get(x))) get() and assign() can be very useful for "computing on the language" issues like this. Bert Gunter Biometrics Research RY 84-16 Merc; /archives/html/s-news/2002-12/msg00030.html (9,496 bytes)
16. Extra Arguments for new plot method ? (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Wed, 06 Nov 2002 15:25:12 -0500; I'm befuddled. ... S-Plus 6.1 R1 for WIndows. I have created a plot method, plot.foo(), for an oldCLass, foo. The method has an additional logical argument, logged, and is defined as: plot.foo<- fun; /archives/html/s-news/2002-11/msg00048.html (7,791 bytes)
17. Re: Extra arguments for plot method (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Wed, 06 Nov 2002 15:45:49 -0500; I found the explanation for my previous query: The method must be defined as plot.foo<-function(foo.obj, ..., logged=F) not as plot.foo <-function(foo.obj, logged=F, ...) to work properly. Question:; /archives/html/s-news/2002-11/msg00049.html (7,789 bytes)
18. Re: A Strange behaviour of "unpaste" (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Sun, 24 Nov 2002 16:15:00 -0500; What version and platform of S-Plus are you using? In S-Plus 6.1 R1 on Windows, it worked with x of length 15 instead of 1500 and Z<-paste(x[1:10],..etc.). However it failed with your example giving; /archives/html/s-news/2002-11/msg00193.html (9,812 bytes)
19. Re: S and macros (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Thu, 03 Oct 2002 08:43:48 -0400; An answer to your question 1 I think is: s.doubleup<-function(...) { out<-list(...) names(out)<-unlist(sys.call())[-1] out } Input is names of your objects; return is named list with your objects as; /archives/html/s-news/2002-10/msg00037.html (13,527 bytes)
20. Re: xlab with apply, hist (score: 1): Author: "Gunter, Bert" <bert_gunter@merck.com>; Date: Tue, 24 Sep 2002 10:44:18 -0400; The apply(hist) behavior appears to be a bug. INSIGHTFUL PLEASE TAKE NOTE! I was able to reproduce it as follows: This appears to be some sort of interaction of apply with hist(). However the labeli; /archives/html/s-news/2002-09/msg00165.html (9,522 bytes)

Current List: 1 - 20
Page: [1] [2] [3] [4] [5] [6]