How to prevent over-fitting?

Question

Anonimo · Accepted Answer

Regularization, test out of sample

Anonimo · Answer

1) get more training data
2) using weight decay to limit capacity of the model
3) using early-stopping 
4) using cross-validation
5) average many different models
6) drop-out

IBM

Domanda di colloquio di IBM

Risposte di colloquio

Aziende seguite

Ricerche di lavoro