commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Phil Steitz <>
Subject Re: [math] Splitting out t- and chi-square tests
Date Sun, 02 May 2004 01:34:28 GMT
Al Chou wrote:
> --- Phil Steitz <> wrote:
>>Currently, o.a.c.m.stat.inference has one interface, TestStatistic, that 
>>includes both t- and chi-square test statistics and tests. I think that it 
>>would be better to split these out into two interfaces -- TTest and 
>>ChiSquareTest (also splitting the impls accordingly).  Any objections to
> Not having any real familiarity with the stat package, here's just a general
> question:  is there any value to TestStatistic remaining as a parent interface
> that TTest and ChiSquareTest extend?  Or would that just be silly?

Interesting idea, but I don't see much that can be abstracted.  Basically, 
TestStatistic provides test statistics and computes p-values for 
Chi-Square and t-tests.   The only thing common among these is that the 
tests are statistical tests and all that could be abstracted would be the 
p-value computation, which is not much IMHO. I am interested in what 
others think, though.

Have a look at the new interfaces after I split them out.  I am going to 
slip support for chi-square tests for independence in two-way tables into 
the chi-square stuff to round out that interface.


> Al
> __________________________________
> Do you Yahoo!?
> Win a $20,000 Career Makeover at Yahoo! HotJobs  
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message