Search String: Display: Description: Sort:

Results:

References: [ +from:dhk@paradox.psych.columbia.edu: 30 ]

Total 30 documents matching your query.

1. Re: Multiple comparisons - general stat question (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Mon, 28 Apr 2003 21:27:19 -0400 (EDT)
a
/archives/html/s-news/2003-04/msg00184.html (7,938 bytes)

2. Re: Generating random real numbers (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Tue, 3 Apr 2001 14:31:37 -0400 (EDT)
I'm a bit confused by the query from Alistair Dempster and the reply from Mike Sawada re random numbers. In the first place, "defined by the normal distribution" is a bit unclear: normal with what me
/archives/html/s-news/2001-04/msg00040.html (9,002 bytes)

3. difference between two predicted vectors (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Fri, 12 Jan 2001 16:46:24 -0500 (EST)
I take it that the 2 forecast vectors differ ONLY in the equations used -- that is, not only the predictands, but also the input data, on which the forecasts are based, are the same. Thus, we are tal
/archives/html/s-news/2001-01/msg00089.html (10,144 bytes)

4. Re: Confidence Interval on Quotient (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Tue, 30 Jan 2001 13:59:04 -0500 (EST)
I've answered this question many times in different places, so I'll just attach my standard answer below: it gives a first and a second order approximation for the coefficient of variation of the rat
/archives/html/s-news/2001-01/msg00225.html (9,809 bytes)

5. Re: [S] uncertainities from ratios (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Wed, 6 Sep 2000 16:36:08 -0400 (EDT)
I happen to have a canned answer to your question about uncertainty of the ratio, from a previous e-mail query (not on s-news). I've expressed the answer in terms of the coefficient of variation (ra
/archives/html/s-news/2000-09/msg00042.html (9,924 bytes)

6. RE: [S] Tests for lack of fit in linear regression (score: 1)
Author: ag.ca
Date: Tue, 29 Aug 2000 10:00:30 -0400 (EDT)
of plots where I need to have two legends in the right margin. The first legend is just a regular legend where the different curves are associated with th
/archives/html/s-news/2000-08/msg00266.html (10,052 bytes)

7. Re: [S] Bootstrapping samples of unequal size (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Tue, 16 May 2000 10:19:24 -0400 (EDT)
t
/archives/html/s-news/2000-05/msg00106.html (9,626 bytes)

8. Re: [S] Statistics question (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Sat, 27 May 2000 09:17:34 -0400 (EDT)
ble
/archives/html/s-news/2000-05/msg00252.html (8,885 bytes)

9. Re: [S] Pr(F) in anova() (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Tue, 7 Sep 1999 14:17:47 -0400 (EDT)
What must be happening in this case is that the added terms decrease the estimated random error, which appears in the denominator of the F tests. The sums of squares and mean squares for old model te
/archives/html/s-news/1999-09/msg00053.html (7,890 bytes)

10. Re: [S] S-PLUS 2000 problem (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Tue, 10 Aug 1999 18:59:04 -0400 (EDT)
I believe that the difficulties you are encountering come from the fact that, as the documentation for data.frame() tells you, "non-numeric vectors are coerced to be factors...". This includes logica
/archives/html/s-news/1999-08/msg00106.html (8,013 bytes)

11. Re: [S] vector of 1:n (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Fri, 16 Jul 1999 09:14:53 -0400 (EDT)
A simple one-liner: rep(1:n, n)[outer(1:n, 1:n, ">=")] will generate the vector you want, 1, ..., n, 2, ..., n, ..., n-1, n, n. Dave Krantz -- This message was distributed by s-news@wubios.wustl.edu.
/archives/html/s-news/1999-07/msg00122.html (7,146 bytes)

12. [S] function syntax (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Sun, 25 Apr 1999 17:48:28 -0400 (EDT)
There are two problems with the function definition The first problem here is conceptual. The second, internal function definition should be just function(x), not (x,y). Since you are applying a fun
/archives/html/s-news/1999-04/msg00225.html (8,683 bytes)

13. Re: [S] Most efficient (best) way to...? (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Tue, 2 Mar 1999 12:13:08 -0500 (EST)
I seldom have occasion to construct a variable of the sort that you want, but when I do, I use the tapply function: group <- tapply(rep(1, nrow(mydata)), list(sex, species)) Ordinarily, tapply() take
/archives/html/s-news/1999-03/msg00016.html (7,976 bytes)

14. Re: [S] Programming Question (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Thu, 7 Jan 1999 13:44:11 -0500 (EST)
I'm not sure what you mean by the "probability of being in one group" but I assume you mean the conditional probability, given the score value. That is, you want to find a critical score x0 such that
/archives/html/s-news/1999-01/msg00038.html (11,692 bytes)

15. Re: [S] attaching & working with a large data.frame (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Thu, 14 Jan 1999 16:50:29 -0500 (EST)
I've worked extensively with quite large data sets in Splus. Generally I find it is best to split the data into a series of smallish data frames, each of which serves some important data-analytic pur
/archives/html/s-news/1999-01/msg00078.html (8,896 bytes)

16. [S] pie charts (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Fri, 13 Nov 1998 10:51:59 -0500 (EST)
I have to confess that I don't remember exactly what Bill Cleveland's arguments were, re pie charts, but I will chime in with a few comments anyway. I don't suppose that my Ph.D. in mathematical psyc
/archives/html/s-news/1998-11/msg00068.html (14,516 bytes)

17. [S] brief replies to the pie-chart critics (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Fri, 13 Nov 1998 12:22:47 -0500 (EST)
I like Type III sums of squares much more than I like pie charts! Every tool can be misused; Why not try to avoid the misuses, rather than throwing out the tools? Robert Garrett's example is pretty c
/archives/html/s-news/1998-11/msg00079.html (9,454 bytes)

18. [S] request for comment about "logical factors" (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Sun, 20 Sep 1998 17:04:39 -0400 (EDT)
In using S+ for exploratory work, I often find myself using table() with one or several of the arguments being logical expressions constructed on the fly, e.g.: table(var1, var2 < 2.5, var3==4) The t
/archives/html/s-news/1998-09/msg00133.html (9,014 bytes)

19. Re: [S] Pi Movie Review(Not work) (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Mon, 3 Aug 1998 08:20:14 -0400 (EDT)
Dear Erin, I guess the mention of the Torah-code bunkum qualifies this as statistics-related: the editors of Statistical Science accepted one of the bunkum articles back in 1994; and perhaps the S la
/archives/html/s-news/1998-08/msg00007.html (8,022 bytes)

20. Re: [S] 95 % Hightest posterior density region (score: 1)
Author: Dave Krantz <dhk@paradox.psych.columbia.edu>
Date: Wed, 19 Aug 1998 19:04:10 -0400 (EDT)
I have a fairly crude function (but it works!) that finds highest density intervals for beta distributions. It is based on the principle that the density must be the same at the upper and lower endpo
/archives/html/s-news/1998-08/msg00139.html (8,269 bytes)


This search system is powered by Namazu