Allow the triggerValue feature of each cluster to be fit to a student-t distribution. This will allow us to automatically determine multi-unit clusters by their confidence intervals overlapping with zero. The reason for choosing student-t over normal is that the former has wider tails than the latter.
Allow the triggerValue feature of each cluster to be fit to a student-t distribution. This will allow us to automatically determine multi-unit clusters by their confidence intervals overlapping with zero. The reason for choosing student-t over normal is that the former has wider tails than the latter.