[SciPy-dev] Homogenizing stats & mstats

Bruce Southey bsouthey@gmail....
Fri Jul 24 21:07:59 CDT 2009


> <snip>
> Great ! That's quite useful. Now, if I may ask for a bit more, what
> would you see in a descriptive.py and in a stattest.py ?
>

Attached is my first attempt at this based on what I started ages ago
so it is probably not current. It is a csv file with columns:

Function Name: name of the function in files

Array: Array function where Stats= stats.py, MoreStats=- morestats.py
or blank if not present

Masked: Masked array function where Mbasic=mstats_basic.py,
Mextra=mstats_extras.py or blank if not present

Support: functions found in the _support.py

Type: In most cases it is obvious where they should go but this column
has my (incorrect) guesses at the possible location:
a) Internal function: small function defined in another function. I
pulled these out in case of duplicated functions but probably should
just disappear.
b) Utility function: Appear to be helper functions that are not
necessary stats based. Some of these probably should have a underscore
to indicate these are not for external use.
c) Summary statistic: Usually provides one or more descriptive statistics
d) Test: Usually involves some sort of testing. I also included the
correlation measures here.
e) Probability function: renamed functions from special that perhaps
should not be defined here.

Comment: Possible actions like depreciate. The question mark indicates
that I thought that the function might be a repeat or variant of other
functions.

If anyone feels I have the wrong location then just say so. Also some
functions may be special cases of others and thus could be
depreciated.

Please correct any errors as there should be some.

Bruce
-------------- next part --------------
A non-text attachment was scrubbed...
Name: stats_functions_24july09.csv
Type: application/vnd.ms-excel
Size: 8730 bytes
Desc: not available
Url : http://mail.scipy.org/pipermail/scipy-dev/attachments/20090724/28d0dd9f/attachment-0001.xlb 


More information about the Scipy-dev mailing list