Mail Archive Search: <+from:parkhurs@indiana.edu>

Results:

References: [ +from:parkhurs@indiana.edu: 31 ]

Total 31 documents matching your query.

1. s (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 15 Jan 2004 21:22:22 -0500; Have a look at the "reverse tests" in Parkhurst DF (2001) Statistical significance tests: Equivalence and reverse tests should reduce misinterpretation. Bioscience 51: 1051-1057. Like Alan, I suggest; /archives/html/s-news/2004-01/msg00086.html (10,644 bytes)
2. Correlated explanatory variables in tree regression (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 2 Jan 2003 15:06:27 -0500; I think I recently read somewhere that, in contrast to multiple regression, tree regression is not negatively affected by correlated explanatory variables. (In fact, what I think I read is that some; /archives/html/s-news/2003-01/msg00003.html (6,689 bytes)
3. Replies to: Correlated explanatory variables in tree regression (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Tue, 7 Jan 2003 13:03:04 -0500; I asked: "I think I recently read somewhere that, in contrast to multiple tree regression is not negatively affected by correlated explanatory variables. (In fact, what I think I read is that some su; /archives/html/s-news/2003-01/msg00019.html (10,532 bytes)
4. Suppressing printing from within a function (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Sun, 29 Dec 2002 14:06:05 -0500; I wrote the following function (based on a message some time ago from Mike rsq.dfp <- function(rtree){ prcp_printcp(rtree) relerr_prcp[,3] 1-relerr[length(relerr)] invisible() } When I call the funct; /archives/html/s-news/2002-12/msg00170.html (7,010 bytes)
5. Re: modeling log transformed values in lme (transformation bias) (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Wed, 16 Oct 2002 20:46:46 -0500; One other caution--if you log transform your data, be careful transforming it the original scale. That can lead to substantial downward "transformation bias," which seems not to be sufficiently well; /archives/html/s-news/2002-10/msg00124.html (11,033 bytes)
6. Printing, or saving to disk, multiple histograms (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Wed, 25 Sep 2002 22:48:58 -0500; I have a dataframe with 20 columns (V1 to V20) and 50 rows. How can I automate printing a histogram for each column, and how can I automate saving those twenty histograms to the hard drive as twenty; /archives/html/s-news/2002-09/msg00186.html (7,087 bytes)
7. Axum graphics vs. s-plus 6 graphics (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 29 Aug 2002 09:16:02 -0500; I've received an offer for MathSoft's MathCAD that includes the Axum 7 graphics package. Does anyone know whether Axum 7 has capabilities (or ease of use) that are not part of the graphics available; /archives/html/s-news/2002-08/msg00192.html (6,518 bytes)
8. Tree regression puzzlement (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Sun, 14 Jul 2002 20:29:50 -0500; I have a dataframe containing a response y and several explanatory x's. If I pairs(), it is clear that y is fairly strongly correlated with at least a couple of the x's. N=16 rows. If I then run rpar; /archives/html/s-news/2002-07/msg00086.html (7,107 bytes)
9. splitting a big scatterplot matrix (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 16 May 2002 21:52:26 -0500; I have a dataframe with 36 columns, and would like to create a scatterplot I just ask for splom(~x), the plots are naturally too small. Is there an easy way to obtain sploms for each sextuple of vari; /archives/html/s-news/2002-05/msg00126.html (6,456 bytes)
10. Trellis sorts conditioning factor. How to prevent? (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 2 Aug 2001 20:14:40 -0500; I have a trellis plot with a formula like y ~ x | factor, where the values of factor are text strings. I want the panels to be plotted in the same order that the factor values are listed in the data; /archives/html/s-news/2001-08/msg00024.html (6,635 bytes)
11. Summary: Controlling order of trellis conditioning factor (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Sat, 4 Aug 2001 17:22:38 -0500; I asked: >I have a trellis plot with a formula like y ~ x | factor, where the values Thanks to the 18 kind folks who suggested solutions. I won't quote them all here (one or two didn't work for me),; /archives/html/s-news/2001-08/msg00039.html (7,127 bytes)
12. Plotting challenge (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 26 Jul 2001 13:17:36 -0500; Suppose I have a dataset like: case before after 1 -1 2.3 2 1.2 2.9 3 3 1 4 2 -1.5 (etc.; N=20 cases) I would like to produce a plot with N vertical bars (possibly touching, or possibly separated wit; /archives/html/s-news/2001-07/msg00203.html (7,129 bytes)
13. Summary: Plotting challenge (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Fri, 27 Jul 2001 15:47:07 -0500; Thanks to the ten kind souls who answered my question, which was: range ... Most proposed solutions involved using segments, arrows, or polygons, and all would have worked. I decided to go with Shawn; /archives/html/s-news/2001-07/msg00215.html (8,821 bytes)
14. Using "by" (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Sat, 28 Jul 2001 14:51:08 -0500; I have a data frame, df say, having four columns, the first of which is a factor variable called, say, fv. I want to us "by" to evaluate a function for each set of rows in df defined by the various v; /archives/html/s-news/2001-07/msg00227.html (6,367 bytes)
15. Need two y's in a trellis panel function (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Sun, 29 Jul 2001 20:57:45 -0500; I have a dataframe, df, of this general form (f is a factor; x, y1, and y2 are numeric): f x y1 y2 1 1 4 6 1 3 2 3 1 5 7 1 2 1 6 2 2 2 3 3 2 4 1 6 3 1 2 4 3 4 3 2 3 5 6 7 I'd like to write a trellis; /archives/html/s-news/2001-07/msg00235.html (6,990 bytes)
16. [S] Using Splus' "layers" (frames and databases) (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Tue, 8 Aug 2000 09:44:04 -0500; was distributed by s-news@wubios.wustl.edu. To unsubscribe send e-mail to s-news-request@wubios.wustl.edu with the BODY of the message: unsubscribe s-news; /archives/html/s-news/2000-08/msg00063.html (8,105 bytes)
17. [S] Reducing splus 2000's selfish behavior with NT4 (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Wed, 9 Aug 2000 13:09:25 -0500; nt.survfit calculates and prints from survfit objects... Is there something more elegant than creating a new function exactly like print.survfit but with "; /archives/html/s-news/2000-08/msg00082.html (8,177 bytes)
18. [S] Do long names slow function execution? (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Wed, 5 Jul 2000 21:04:10 -0500; I generally like to use LongDescriptiveNamesLikeThis for variables when I'm programming. Because s-plus is not compiled, would that slow execution of a function (in a large simulation), compared to s; /archives/html/s-news/2000-07/msg00024.html (7,346 bytes)
19. [S] Summary: Do long names slow function execution? (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Sat, 8 Jul 2000 08:14:54 -0500; Recently I asked: I generally like to use LongDescriptiveNamesLikeThis for variables when I'm programming. Because s-plus is not compiled, would that slow execution of a function (in a large simulati; /archives/html/s-news/2000-07/msg00046.html (8,243 bytes)
20. [S] Error with dotplot (score: 1): Author: "David Parkhurst" <parkhurs@indiana.edu>; Date: Thu, 15 Jun 2000 13:45:48 -0500; When I try to execute the trellis dotplot example from the help in splus 2000, rel. 2, under NT4.0, I get the following result: dotplot(variety ~ yield | year * site, data = barley, aspect = 0.4, xla; /archives/html/s-news/2000-06/msg00102.html (7,327 bytes)

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