Introduce shuffle trick to get more local P-values

[vbcwonderland]

Sampling by shuffle threshold and get P Values for each shuffle group. The result pVals_list will then being used in other methods.
For example: generate each of its own coresponding Anderson-Darling Test P Value.

[jdramsey]

Should this: int shuffleTimes = (int) Math.ceil(1 / shuffleThreshold);
be this?: int shuffleTimes = (int) Math.ceil(sampleSize / shuffleThreshold);

for some sample size?

[vbcwonderland]

Hi Joe,

I think it should be (int) Math.ceil(1 / shuffleThreshold) because the shuffleThreshold here is a decimal, e.g. 0.3.
Then in order to cover all samples at least statistically for once (i.e. no shuffle, use the whole sample), we should do int) Math.ceil(1 / shuffleThreshold) so that it can achieve the same. e.g. 1 / 0.3 = 3.333 -> round up to 4 times, would make sure if we take 30% of our current data for 4 times, we would expect to use all data points for once.

Best,
Verity

[jdramsey]

Oh i see what you mean... let me go over this again now.

[jdramsey]

Looks good.