A decision threshold is the cutoff value used to convert a model score or probability into a discrete decision or class prediction. In binary classification, changing the threshold can alter the balance between false positives and false negatives.
Synonym for classification threshold.