Shamus Husheer
Shamus Husheer
See also https://github.com/scipy/scipy/issues/14472
I think this category of solution comes under my proposed solution (1), "modifying x-values or combining close points, for some definition of close (might be very data-specific)". For this particular...
Ok, I had a play with this, and think I have a more elegant solution, which could be more widely useful. The basic idea is to compress the representation of...
Good investigative powers ;) So I think what needs to be done is the pattern for CubicSpline should be followed, where there is a bc_type parameter on creating the spline....
By the way, the big jump at the boundary in your example code above is due to failing to match the endpoint conditions in the input data. Insert this line:...
Whilst you're modifying _make_spline() you might want to look at this error path: if not all(dx > 0): raise ValueError ... There's no good reason for this to break the...
I assume it will take you some time to look into the correct boundary conditions for periodic functions. In the meantime, here is a function that applies my earlier "compression"...
> what is the theoretical justification for the chosen pvalue? I don't choose a fixed pvalue, I calculate the probability that a sample size of N is all the same,...
There are a few failures on tests, but these seem entirely unrelated to the runs test, so I presume reflect other possible issues in the wider package. I have therefore...
> Out of curiosity, How did you run into this problem with only one run? If the threshold is mean or median, then we need to have at least 2...