Returning Interval instead of tuples in distributions' Support members
#301
Conversation
Better fitting Support member without the superfluous lambda. Marking the other Support member as obsolete.
... of NegativeBinomial
There was a problem hiding this comment.
It seems like the support of the F distribution is dependent on the degrees of freedom used (dof1). At least here it says that its either [0,inf) or (0,inf) depending on dof1 beeing 1 or not. Do you have some thoughts if its correct to match dof1 beeing 0, or 1 respectively and then return different intervals?
There was a problem hiding this comment.
Good point. Additionally, it seems that dof2 is irrelevant for determining the value of Support, so this parametert can be removed too.
There was a problem hiding this comment.
I assume the description is an artifact from the exponential distribution (basically it says so in the /// comment) and the interval should be (0,inf). The same may apply to LogNormal
Reference
There was a problem hiding this comment.
Because interval is -infinity, the interval should be open. Sorry about the misleading comments above each distribution support. Could you exchange exponential with the appropriate distribution discription for each distribution? If not, no worries I can take care of if, just let me know.
There was a problem hiding this comment.
Sure. Sorry, it seems like I missed this one. But you're right, some of the current comments are not helping 😄
Codecov Report
@@ Coverage Diff @@
## developer #301 +/- ##
=============================================
+ Coverage 46.31% 46.61% +0.29%
=============================================
Files 148 148
Lines 16049 16182 +133
Branches 2176 2193 +17
=============================================
+ Hits 7433 7543 +110
- Misses 7968 7986 +18
- Partials 648 653 +5
|
Removed `dof2` as a parameter of `Support`, since it plays no role there. Also added a helper module to the F distribution file with a helper pure function. I hope it's ok.
Supportmember for all probability distributions #292This pull request introduces breaking changes.
Although these are just some small changes, they do represent breaking changes as the public contract of the
Supportmembers are changed.The
Supports of the exponential distribution and of the discrete negative binomial have also been fixed (were previously closed intervals, but they should have the right side "open").Check list
The project builds without problems on my machine
Adapted unit tests regarding the added features