Shan Huang
Shan Huang
Yes the "dynamic threshold" is based on information fraction, which is ratio of current sample size and estimated sample size for the experiment. Here is the method we use: https://github.com/zalando/expan/blob/master/expan/core/early_stopping.py#L24-L36
Whether it is day-by-day analysis or other periods, will depends on how your code calls ExpAn.
Hi @liuweizzuie, I did a quick search, this seems to happen with Windows with default Chinese encoding with Python3. See a same discussion here: https://github.com/sshwsfc/xadmin/issues/330. If you have some time,...
@CHI-MING-LEE thank you!
hi @marcosmarxm, thanks for reaching out! No, there is no specific point of using `stan`. :) We started with it three years ago. Back then our developer was most familiar...
@aaron-mcdaid-zalando works for me. It could be tox cached an old version of pystan for you. Try delete .tox directory and run again. :)
Yes this is about result structure improvement. The two points might not make sense. :) Just a reminder for points to discuss.
One idea we've touched is to create the ```Result``` class with mandatory and optional fields that is used to hold results for different types of methods.
thanks @andompesta for the contribution! The test failed after the new pystan version because we hard coded the numerical test values before, which is not right way to test. I...
@jbao Is there a reason you use z-test(instead of t-test as Darya described above) when deciding early stopping in the group sequential method? :)