Hyperparameters / Algorithm Choice


We have learnt in this specialization about many choices that we can choose from in order to build an effective model.

However, I don’t think we got deep into how to choose one over another.

For example, for Regularization, when do we choose drop-out, L1 or L2? Or, how can we make an informed decision about that?

For the optimization Algorithms, when do we choose ADAM, RSPBM, GD or GD with Momentum?

Have you seen this link ?