Please sign in to comment.
turns out matrix has to explicitly call its constructor "__construct(…
…)" to work When the constructor was "matrix()" everything worked fine until we moved to the namespace in SparkLib, then the constructor wasn't being called. This was a silent failure that popped up in Regression by complaining about undefined indexes in the coefficient array. Before this fix all polynomial regression was effectively nonfunctional.
- Loading branch information...
Showing with 2 additions and 1 deletion.