Becoming a superforecaster, decision science for better human predictions - Pavel Atanasov-the data scientist show#036
Daliana's Game
00:00
Is 50 50 the Default for a Binary Classification?
The default of the motto always calebry at point five. But that may now be true for different cases, right? I think exam as the data signed is, we need to think about maybe your mottos kind of under confident in certain cases. Maybe you should dicotomize it at, i dont know, point four or somewhere, depends on the use. This is very important also for our machine learning moto. Think about a binary classification. He ye, i think in most cases, 50 50 is default. And forleck, if it's a bindery, like, yes now, but it's really the best one. Lhik, usually you know
Transcript
Play full episode
