Maksim Terpilovskii comments

Results 30 comments of


                                            Maksim Terpilovskii

Dunnetts test

@comatrion Thank you for suggestion! I will implement it soon.

@mysticaltech This test relies upon multivariate Student's t-distribution function ([pmvt](https://www.rdocumentation.org/packages/mvtnorm/versions/1.1-0/topics/pmvt) in R). AFAIR, I could not find a usable implementation of it in Python a year ago. It would be...

Dunnetts test

I have looked at the code in `pmvt` R package and even tried to implement bindings to Fortran code, but failed. This is not trivial. No progress so far, unfortunately....

Dunnetts test

@Ronny-PDT Thank you for your interest! Unfortunately, I do not have enough time to work with it now. You can try to implement it yourself and do a pull request

Dunnetts test

@Tsuchihashi-ryo Thank you very much!

Question: Post-hoc dunn return non-significant

I have checked the algorithm and found no errors. Dunn suggests ranking all data in the [original paper](https://www.stat.cmu.edu/technometrics/59-69/VOL-06-03/v0603241.pdf).

Solving ValueError; 'All numbers are identical in mannwhitneyu'

I have rechecked this with the latest version of `scipy`. Now it is not throwing such an error. No fix is needed.

posthoc_tukey yields p-values of 0.900 and 0.100 which are also different to scipy.stats.tukey_hsd

Hi! Thank you for reporting this. You should be using `posthoc_tukey_hsd` function instead of `posthoc_tukey`. But I do not like how both of them are implemented. Scipy's function was added...

Alpha cannot be specified for critical difference diagram

Hello! Thank you for pointing at this. This test was based on statsmodels' implementation of Studentized range distribution. It is bound to 0.1, there is no way to specify alpha...

Fix overlapping crossbars

Thank you! I'm gonna review all the related code in your pull request and this package