Hi Stephane & Didier,
at the moment i am on holiday,
enjoy :)
but i noticed the nice & interesting
developments in scismalltalk. regarding further statistical
tests: in most cases they simply calc the probability of the
0-hypothesis. hence for a simple common interface it would
perhaps make sense to have an abstract object for these things,
that has accessors for the data, the test-statistic, the p-value
and a #rejectEqualityHypothesisWithAlpha:. and for
2-sample-tests also a second data-var. and perhaps a common
output format for the result (so that one gets an idea how to
structure the output if one implements a new test). eventually
it could make sense to have a short look at the abstract
KolmogorovSmirnov object which does some of these things more or
less for kstests, so - if one has such an abstract object - one
can also replace #KolmogorovSmirnov with that object.
Yes for dummy like me it would be good :)
As we were discussing in the sushi shop :), the library should
encapsulate the know-how so that
math-naive like me can use and if they want open the box.
Hi Vincent,
it would perhaps be helpful if you could specify your proposal
"Standard statistical tests: normality, variance, mean,
distribution" a bit, because eventually some of these are
already implemented and would just need a wrapper to be easily
usable. perhaps it would be enough if one would know which
R-functions you use or so, and what kind of return you'd expect.
werner
--
You received this message because you are subscribed to the Google
Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
[hidden email].
For more options, visit
https://groups.google.com/d/optout.
Locked
