Hi Serge,

i need to simplify some code in dhb-numerical and i'd like to know whether i may in the process of doing that move your DhbExponentialDistributionTest to Math-Tests-DHB-Numerical? if you put it in Math-DHB-Numerical in order to refind it simpler, i could put it in a separate protocol tests-statistics in Math-Tests... , would that be ok with you?

btw i looked at the test and i probably dont understand the comment there. i often make inline comments just for myself that are <g> ununderstandable for other users, hence i usually just skip comments i dont understand, im just not sure whether this "testing a random sample seems suspect" is a comment like that. i looked at that test:

"you have:"

eg := DhbExponentialDistribution scale: 10.

"with mean & stdev"

eg average." 10"

eg standardDeviation ." 10"

"a sample mean follows a normal distribution with a stdev of stdev/(n sqrt) :"

10.0 / 10000 sqrt." 0.1"

"you test for an interval of mean 10 +-0.5, hence +-5*stdev"

nd:=DhbNormalDistribution new." Normal distribution( 0, 1)"

"hence the probability of a test failure is around (well, the book says the error in the [first part of the] calc is smaller than 7.5e-8):"

(nd distributionValue: -5)*2. "5.742099990047038e-7"

even if you add twice the possible error this is not really high <g>, but, well ok, what seems suspect?

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.