Avoiding Overfitting :
However remember there that in the previous lecture, there is over fitting that common problem in machine learning. Furthermore details to decision trees suffer from this is because they are trained to stop where they have perfectly classified all the training data that i.e., each branch is extended that is far enough to correctly categorise the examples relevant to that branch. In fact many other approaches to overcoming overfitting in decision trees have been attempted but as a summarised by Tom Mitchell there these attempts fit into two types as:
• Just stop growing the tree before it reaches perfection, and • Now allow the tree to fully grow so then post-prune some of the branches from it.
Hence the second approach has been found to be more victorious in practice. Means that both approaches boil down to the question of determining the correct tree size. Here you can see Chapter 3 of Tom Mitchell's book for a more detailed description of overfitting avoidance in decision tree learning.