commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Phil Steitz <stei...@yahoo.com>
Subject Re: [math] RandomData and ValueServer Failures . . .
Date Tue, 17 Jun 2003 15:12:52 GMT
--- Phil Steitz <steitzp@yahoo.com> wrote:
> 
> --- "Mark R. Diggory" <mdiggory@latte.harvard.edu> wrote:
> > 
> > 
> > Al Chou wrote:
> > > --- "Mark R. Diggory" <mdiggory@latte.harvard.edu> wrote:
> > > 
> > >>I appear to "occasionally" get JUnit test failures from ValueServer and

> > >>RandomData Tests. This would appear to be because the mean sampled 
> > >>values can sometimes deviate from the expected mean even for 1000 case 
> > >>draws, I know this happens "rarely", just enough over the last month or

> > >>so for me to start to notice this behavior. Oldly, when its off, its off

> > >>in a big way, so its not just a matter of changing the tolerance.

One more remark on this.  My previous response explains the RandomDataTest
failures -- these are to be expected roughly 1 in every 1000/(number of tests)
timess.  The ValueServer is more complex (at least the testNextDigest test). 
What is going on there is that first the test is "digesting" an input file to
compute an empirical distribution describing the data.   This is done by an
EmpiricalDistribution instance.  The technique used divides the data range into
"bins", computes bin frequencies and within-bin means and variances and then
stores this information for use in generating data.  The default bin count is
set at 1000 and the ValueServer tests do not override this.  This is a mistake,
since the test file has only 1000 records in it. (Generally the bin count
should be an order of magnitude less than the number of records, as indicated
in the javadoc ;-)) I should have changed this when I cut the test file down to
1000 records.  This may be what is causing the relatively high incidence of
failures for the ValueServer tests.  I will submit a patch decreasing the bin
count for these tests when I get back at the end of this week.  Thanks for
pointing this out.  Strangely, it hasn't happened to me (i.e., the ValueServer
tests failing).

Phil


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org


Mime
View raw message