Mail Archive Search: <+from:john.maindonald@anu.edu.au>

Results:

References: [ +from:john.maindonald@anu.edu.au: 42 ]

Total 42 documents matching your query.

1. Re: log(-log x) distribution (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Tue, 20 Mar 2001 08:25:05 +1100 (EST); Do you have Xi = p, or Xi=1-p, where p is a proportion/ probability. With Xi=1-p this is a complementary log-log distribution. It can be motivated as an extreme value distribution. The idea is that f; /archives/html/s-news/2001-03/msg00148.html (7,952 bytes)
2. Re: [S] uncertainities from ratios (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Thu, 7 Sep 2000 09:55:55 +1100 (EST); Note also the simplified version of this formula, but allowing for correlation between A and B: var[R] = 1/B^2(R^2 V22 - 2 R V12 +V11) where var[A]=V11, var[B]=V22, cov[A,B]=V12. Note also that an ex; /archives/html/s-news/2000-09/msg00044.html (11,451 bytes)
3. Re: [S] NA's in acf (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Tue, 15 Aug 2000 13:55:58 +1000 (EST); ope this helps! Sincerely, Erin M. Hodgess, Ph.D. Associate Professor Department of Computer and Mathematical Sciences University of Houston - Downtown One; /archives/html/s-news/2000-08/msg00140.html (7,417 bytes)
4. Re: [S] logit regression: partial residual plot (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Mon, 17 Jul 2000 09:33:18 +1000; The plot that you seem to describe is a component plus residual plot, which one can get with plot.gam(.., residuals=T) With non-normal errors there are major issues in deciding what residuals it may; /archives/html/s-news/2000-07/msg00107.html (10,968 bytes)
5. Re: [S] Transfer between R and S-PLUS (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Thu, 22 Jun 2000 10:02:48 +1000 (EST); .... It depends on what one is doing. If the use of R and S-PLUS does not go much beyond writing functions for one's own use, my experience has been that it is pretty easy to make the transition, in; /archives/html/s-news/2000-06/msg00154.html (9,058 bytes)
6. [S] Fitting mixtures of exponentials (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Wed, 19 Apr 2000 12:39:30 +1000 (EST); I'd be grateful for references to software for this, and/or for commentary on likely numerical difficulties? The chief challenge, I guess, is to get good starting values for maximising the likelihood; /archives/html/s-news/2000-04/msg00163.html (7,428 bytes)
7. [S] Handling of missing values in table() (score: 1): Author: john.maindonald@anu.edu.au; Date: Thu, 02 Mar 2000 15:13:56 +1100; I am using S-PLUS2000 (Release 2) under Windows 98. The following is weird. It is also very annoying, as I do generally want numbers of NAs noted when I do a tabulation on data that has been imported; /archives/html/s-news/2000-03/msg00018.html (10,040 bytes)
8. Re: [S] Handling of missing values in table() (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Fri, 3 Mar 2000 09:31:50 +1100 (EST); Thanks to Brian Ripley, Judi Scheffer, Mike Hoekstra (who reported an exchange with S-PLUS Tech support), Anne York, Don McQueen and Bill Dunlap for their replies. It may be helpful to others who are; /archives/html/s-news/2000-03/msg00034.html (9,719 bytes)
9. [S] Handling of NaNs (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Fri, 3 Mar 2000 09:45:27 +1100 (EST); I forgot to say that Brian Ripley sent me a later correction - dump() sometimes NaNs into NAs. As I noted, the evidence is that import.table() in S-PLUS 2000 creates a bit pattern which is neither Na; /archives/html/s-news/2000-03/msg00035.html (7,122 bytes)
10. [S] Anomalous behaviour when importing from Excel 2.1 files (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Wed, 8 Mar 2000 15:29:27 +1100 (EST); Here's another to add to the list. When importing numeric data that has come from Excel 2.1 on the Macintosh, blanks in the first row get replaced by "?", and the whole column is imported as a factor; /archives/html/s-news/2000-03/msg00077.html (7,532 bytes)
11. Re: [S] lme and gls; Function for extracting Least-squares means (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Fri, 07 Jan 2000 17:47:12 +1100; .... Also, model.tables() appears dangerous. It requires a warning that it gives meaningful output only for a limited class of models (Which class?) The following is a balanced incomplete block desig; /archives/html/s-news/2000-01/msg00035.html (9,829 bytes)
12. [S] Printing from the MathSoft web site (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Thu, 4 Nov 1999 16:52:13 +1100 (EST); I use Netscape Communicator 4.07 on a PC running Windows 95B. When I printed out the S-PLUS2000 Release 2 README file, from Netscape, it printed out a huge number of pages with 1 line per page. I enc; /archives/html/s-news/1999-11/msg00031.html (7,977 bytes)
13. RE: [S] Printing problem with MathSoft website resolved (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Wed, 10 Nov 1999 08:36:07 +1100 (EST); The matter does however draw attention to serious communication issues. It had been there at least since the 13th September (in the file migrating.htm), and almost certainly for quite a while longer.; /archives/html/s-news/1999-11/msg00095.html (9,497 bytes)
14. mmary: C/C++ question (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Tue, 30 Nov 1999 09:47:19 +1100 (EST); First of all you have to say what you mean by A:C when C and/or A are not present. One does this by placing constraints on the parameters which specify the model which includes main effects. It is th; /archives/html/s-news/1999-11/msg00283.html (10,583 bytes)
15. Re: [S] Re: Excel Import Problem (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Wed, 29 Sep 1999 08:52:52 +1000 (EST); Are you using S-PLUS4 or 4.5, or 2000? In version 4.5 it appeared that the contents of the first row determined the mode for the whole column. It ignores (at all events with Excel 97) rows beyond num; /archives/html/s-news/1999-09/msg00270.html (9,633 bytes)
16. Re: [S] Default settings for trellis (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Fri, 14 May 1999 09:51:34 +1000; .... You can control the height of the strip text (and the height of the strips) with something like: par.strip.text=list(cex=0.8) As for (2), this comes up frequently. Several of us have our own str; /archives/html/s-news/1999-05/msg00156.html (10,950 bytes)
17. Re: [S] Least Square Means in Splus found it BUT ... (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Sun, 30 May 1999 10:40:30 +1000 (EST); My understanding from the earlier discussion was that there was a perception, whether well-founded or not, that the FDA requires (maybe only sometimes) Type III sums of squares. It would be a service; /archives/html/s-news/1999-05/msg00323.html (10,264 bytes)
18. Re: [S] Testreport: Math. & stat. software for data analysis (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Wed, 21 Apr 1999 08:46:51 +1000 (EST); ..... I am concerned that the comparisons should do more to distinguish the trivial or archaic from what is consequential. A quick scan reveals the following: 1) An identify matrix (2.2) would be tri; /archives/html/s-news/1999-04/msg00183.html (9,721 bytes)
19. [S] S-PLUS4.5 import from Excel truncations (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Fri, 30 Apr 1999 15:55:24 +1000 (EST); I'm using Splus4.5, release 2, under Windows 95. Some time ago some documented the truncation of text columns from Excel spreadsheets to 31 characters. Now I find that the dataset likewise gets trunc; /archives/html/s-news/1999-04/msg00290.html (7,452 bytes)
20. Re: [S] model selection in glm - logit link vs identity (score: 1): Author: John Maindonald <john.maindonald@anu.edu.au>; Date: Thu, 28 Jan 1999 08:54:21 +1100 (EST); Brian Ripley responded One does have to worry about the difference in b0. How do the fitted values for the three fits compare? I doubt that this is a convergence issue, because it ought to affect oth; /archives/html/s-news/1999-01/msg00186.html (10,768 bytes)

Current List: 1 - 20
Page: [1] [2] [3]