Discrete data fit
koko12131 opened this issue · 5 comments
Thanks a lot for this useful tool. But if my data is not continuous but discrete, ex Integers, what should I do ?
Checkout this question/answer on stackoverflow:
https://stackoverflow.com/a/62339711/13730780
I think this was a similar questions as yours. The link refers to a answer where distfit is used to solve the issue.
Let me know wether this helps!
thank you for your very quick reply. The Q/A really helped me. But there is another question about the Q/A in which a parameter-"smooth=10" was used to work with integers. My question is how to determine the value of parameter-"smooth"?
This seems to depend on the number of samples that is being used.
I restructured the sphinx pages, and included a new section to examine integer fitting and the effect of the smooting parameter.
For both continuous and integer values, the conclusion is that the smooting window seems only to be usefull with very small samples sizes. Check out the sphinx pages for code and figures that support this conclusion.
https://erdogant.github.io/distfit/pages/html/Performance.html#integer-fitting
thank you so much for your constructive comments!this does help me a lot!
For future references; https://erdogant.github.io/distfit/pages/html/Discrete.html