ficerr

Just wondering if anyone has got any updates or ideas on resolving this issue, would be great to know.

@jcrudy Yes, it is an issue as the difference in the results from one PC to another is dramatic, it is not like a rounding error in the 10th decimal...

@jcrudy Across different computers: basis functions are the same, knots are the same (at least up to 3rd decimal place, I do not know where to find the exact values...

@jcrudy Well, the effect of that instability depends on how sensitive you want your model to be. Regarding the differences across computers: for instance, sometimes I get a very small...