Suppose you have a returns vector r, which contains values like +3,
-2, -7, +3, etc. Then say:
require(tseries);
test = runs.test(factor(sign(r)))
print(test)
If you want to access the prob value of the test from within a
program, this is found in test$p.value
--
Ajay Shah
Hi Mohammad,
There an implementation of Knuth's run test in the pLab programs. While these
are apparently not being maintained anymore, there is a good description of
the code here:
http://random.mat.sbg.ac.at/team/#software
Download this file - the Mathematica code is on page 65.
Leeb, H. a
Dear Andy,
Usual literature of runs:
Run: a run is defined to be a maximum subsequence of
like elements.[Mendenhall, Scheaffer, and Wackerly
(1986), Mathematical Statistics with Applications, 3rd
Ed., Duxbury Press, CA, Page 597]
Run: a sequence of 1 or more identical symbols that is
preceded an
I don't know the literature at all, but intuitively, you would need an
ordering relation to be able to define `runs'. How would one do that in
more than one dimension?
Cheers,
Andy
> From: WilD KID
>
> Dear all,
>
> Does R or S-plus or any of their packages provide
> Non-parametric "Run test"
Dear all,
Does R or S-plus or any of their packages provide
Non-parametric "Run test" (which tests whether a
sequence of numbers might be random or not)? If yes,
i'd like a numerical illustration of this test.
Any response / help / comment / suggestion will be
greatly appreciated. Thanks in adva