Choosing a classifier’s yes/no cutoff: A post asks what fits your use case
Most classifiers output probability estimates, such as “70% likely,” a post explains. Turning those estimates into yes/no decisions means choosing a decision threshold.
TLDR
A post asks how to choose the right decision threshold—the cutoff used to turn a classifier’s probability estimate into a yes/no decision. Its central question is how to know which threshold fits a particular use case.
Combined views
1.3K
2 Sources, first seen 23d ago
5 likes