- 1. [S] Re: seq and floating point (was: a meaningless subject line) (score: 1)
- Author: Tim Hesterberg <THesterberg@statsci.com>
- Date: Fri, 18 Feb 2000 09:07:14 -0800 (PST)
- As Brian Ripley and Michael Prager pointed out, this occurs because of floating point arithmetic. Note that seq(from=a, to=b, by=c) is not guaranteed to have the last value approximately equal to `b'
- /archives/html/s-news/2000-02/msg00169.html (9,084 bytes)
- 2. [S] Re: seq and floating point (was: a meaningless subject line) (score: 1)
- Author: Tim Hesterberg <THesterberg@statsci.com>
- Date: Fri, 18 Feb 2000 09:21:21 -0800 (PST)
- That should be: seq(from=a, to=b, length=round((b-a)/by) + 1) Tim Hesterberg -- This message was distributed by s-news@wubios.wustl.edu. To unsubscribe send e-mail to s-news-request@wubios.wustl.edu
- /archives/html/s-news/2000-02/msg00171.html (7,676 bytes)
- 3. Re: [S] Generalized GAM (score: 1)
- Author: Tim Hesterberg <THesterberg@statsci.com>
- Date: Mon, 21 Feb 2000 12:12:04 -0800 (PST)
- Trevor Hastie provided a wonderful solution to problems like this -- to use regression splines, e.g. bs() or ns() (rather than s() or lo()). E.g. lm(Z ~ bs(X) + Y * bs(X)) You can specify the degrees
- /archives/html/s-news/2000-02/msg00192.html (11,286 bytes)
- 4. Re: [S] stacking data.frame (score: 1)
- Author: Tim Hesterberg <THesterberg@statsci.com>
- Date: Sat, 26 Feb 2000 09:49:04 -0800 (PST)
- Equivalently (but simpler): tmp[rep(1:nr, each=3), , drop=FALSE] The drop=F part is necessary if tmp had a single column. Alternately, set the dup.row.names attribute to non-NULL, e.g. one of: attr(
- /archives/html/s-news/2000-02/msg00273.html (9,513 bytes)
This search system is powered by
Namazu