Re: st: Features for Stata 14

[László Sándor <sandorl@gmail.com>]

Simpler things: subsampling analogue of bootstrapping.

Thinking outside the box: automatic recommendations for more efficient
coding, esp. with big data. Stata is a high-level language for many
amateur programmers, but many CPU hours and RAM (or as a consequence,
quite some grant money) is wasted on running code that the author did
not know they could write better. Automatic improvements are tricky,
for sure. But suggestions only take some parsing of do files, maybe?
And they could always be turned off.

My projects are seriously hindered by my shared resources being slowed
down by other users' code running for days and with hundreds of GBs of
memory, and the friendly research environment won't enforce rules
while it cannot offer advising on the code either.

E.g. Dan's collection is a strong start, esp. with Joe's slides linked
in at the end. Plus some advice on trying to avoid commands that
preserve or sort. And the tips from the journal, of course.
http://www.nber.org/stata/efficient/

Another major thing probably not changing anytime soon: data-access
performance. http://www.asdfree.com/2013/03/column-store-r-or-how-i-learned-to-stop.html

On Tue, Sep 3, 2013 at 11:22 AM, Ariel Linden, DrPH
<ariel.linden@gmail.com> wrote:
> I would really like to see a suite of data mining/machine learning analytic
> tools, such as classification and regression trees, etc.
>
> Date: Mon, 2 Sep 2013 10:00:28 -0500
> From: William Buchanan <william@williambuchanan.net>
> Subject: st: Features for Stata 14
>
> Since it is still fairly early in the development cycle for the next release
> of Stata, I thought it might be good to start a thread about things that
> people would like to see added to the next release.
>
> I would definitely be interested in seeing some updates/expansions to the
> graphics capabilities of Stata.  Alpha level blending is something that has
> come up several times and adding interactive graphics would be a great
> addition to the existing commands (e.g., functionality that is common in
> Tableau and several packages in R).
>
> Documentation of the lower level graphics commands to make it easier for
> users to extend the graphics capabilities.
>
> Latent Class Analysis and Latent Transition Analysis with support for all
> types of manifest variables.
>
> Multivariate mixed-effects models
>
> Exploratory Factor Analysis with nominal/ordinal/non-normal variables
>
> Improved debugging tools (e.g., stepwise evaluator for programs, more
> informative error messages, etc...)
>
> New ways to generate samples of data with specified correlation
> structures/distributions
>
> I'm curious to see what other people have to say on this topic as well.
>
> Billy
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/