How to measure how a decision tree generalises? Train, validation, test sets

Thankyou !!! any idea for this? How to find a target attribute in a test set using a decision tree or Random forest or SVM classifier built from the training set?