# How can generalization error be estimated?

How would you estimate the generalisation error? What are the methods of achieving this?