sdl@linus.mitre.org (Steven D. Litvinchouk) (11/22/90)
I was wondering if anyone could point me to any sources of information and experience, in using APL for exploratory data analysis. What I mean by that is analyzing multivariate data (including, but not limited to, time series data), when the trends and models that may be present in the data are unknown; the goal is to discover those trends and models. Also, if there are any APL toolkits for doing this (e.g. scatterplot brushing), I would like to know that too. Also, I had heard that there had been work on developing object-oriented extensions to APL. Can anyone provide me with any pointers to that work? Thanx in advance! Steven Litvintchouk MITRE Corporation Burlington Road Bedford, MA 01730 (617)271-7753 ARPA: sdl@mbunix.mitre.org UUCP: ...{att,decvax,genrad,necntc,ll-xn,philabs,utzoo}!linus!sdl "Where does he get those wonderful toys?" -- Steven Litvintchouk MITRE Corporation Burlington Road Bedford, MA 01730 (617)271-7753 ARPA: sdl@mbunix.mitre.org UUCP: ...{att,decvax,genrad,necntc,ll-xn,philabs,utzoo}!linus!sdl "Where does he get those wonderful toys?"
rdnelson@phoenix.Princeton.EDU (Roger D Nelson) (11/23/90)
In article <SDL.90Nov21172049@lyra.linus.mitre.org> sdl@linus.mitre.org (Steven D. Litvinchouk) writes: > >I was wondering if anyone could point me to any sources of information >and experience, in using APL for exploratory data analysis. What I >mean by that is analyzing multivariate data (including, but not >limited to, time series data), when the trends and models that may be >present in the data are unknown; the goal is to discover those trends >and models. Also, if there are any APL toolkits for doing this (e.g. >scatterplot brushing), I would like to know that too. STSC sells STATGRAPHICS, an APL based statistical and graphic analysis package. It is necessary (in my opinion) to run it within their APL program, although STATGRAPHICS is standalone, in order to have access to APL for flexibility in exploratory data analysis. I use the package regularly for some of the purposes you mention, and find it satisfactory. While not as broad as SAS or BMDP it does have multivariate regression, including some nonlinear fitting capability (you supply the algorithm or a reasonable guess), good time series functions, some categorical analysis, many distribution functions, and virtually all canonical statistics, plus well integrated graphics. I don't have any connection with the company, and indeed find them frustrating, but I do get a lot of work done using STATGRAPHICS. Roger Nelson rdnelson@phoenix.princeton.edu