Open simonbyrne opened 5 years ago
One option is to use corrected = 1
/2
/3
, corresponding to the types from the Joanes and Gill paper. The problem is that this doesn't match up with the current boolean values (false => 1
, true => 3
).
The problem is that this doesn't match up with the current boolean values
I'm not sure what you mean. We don't pass boolean values to skewness
or kurtosis
right now as far as I can tell.
ah, you're right, I was thinking of std
.
One option is to use
corrected = 1
/2
/3
That seems reasonable to me. Another option would be something like
@enum BiasCorrection type1 type2 type3
then use correction=type1
/type2
/type3
.
Using integers sounds fine. We could also use symbols, which are more common than enums in our API (that would avoid exporting type1
).
skewness and kurtosis have multiple ways to do biasedness corrections, e.g.
How should we specify these?