ecacarva
Hello, everybody.I would like to know the best way in how to evaluate the quality of a BN when the structures and parameters are learned from data.Some good references would be good.Thank you
Quote 0 0
Dan
There are 2 contexts:- Unsupervised Learning: you can evaluate the representation quality of the joint probability distribution with Analysis | Network Performance | Overall- Supervised Learning: Analysis | Network Performance | Target allows to measure the quality of the network for predicting the target variableNote that you can also used the cross-validation validation tools (Tools | Cross Validation)
Quote 0 0