What is a clipped and uncircumcised tree in the Age?

Question

In decision tree J48, when we say that the tree is truncated or not, what is the difference?

Answer 1

Maintenance free trees anymore. What happens is that basically the tree is created in accordance with the implemented algorithm, and if pruning is turned on, an additional step shows which nodes / branches can be deleted without affecting performance too much.

The idea of ​​pruning is that, in addition to simplifying the understanding of the tree, you reduce the risk of reassigning training data. That is, having the ability to classify training data is (almost) excellent, but nothing more, because instead of studying the basic concept, the tree recognized properties that are inherent and specific to the training data.

Answer 2

I would like to add the following to Lars's answer. Adapted from the following link

Many algorithms try to “trim” or simplify their results. Cropping gives less, more easily interpreted results. More importantly, pruning can be used as a tool to correct potential retraining ....

J48 uses two cropping methods.

The first is known as replacing a subtree. This means that the nodes in the decision tree can be replaced by a sheet - basically reducing the number of tests on a certain path. This process begins with the leaves of a fully formed tree and runs back to the root.

The second type of trim used in J48 is called a subtree. In this case, a node can be moved up to the root of the tree, replacing other nodes along the way. Increasing subsamples often has a negligible effect on decision tree models. Often there is no clear way to predict the usefulness of an option, although it may be advisable to try to disable it if the induction process takes a long time. This is because raising a subtree can be somewhat computationally complex.

Answer 3

Given that Weka is a machine learning kit, this is similar to what they mean:

http://en.wikipedia.org/wiki/Pruning_ (decision_trees)

In short, pruning a decision tree is the elimination of possible solutions that do not bring much benefit.

However, I have not used weka and am not familiar with this. Relate to the other answers and see if what they say makes sense first.