Julia for Data Science by Unknown

Author:Unknown
Language: eng
Format: epub, mobi
Publisher: Packt Publishing

Disadvantages of decision trees

We've covered where decision trees are suited and their advantages. Now we will go through the disadvantages of decision trees:

There is always a possibility of overfitting the data in decision trees. This generally happens when we create trees that are over-complex and are not possible to generalize well.

To avoid this, various steps can be taken. One method is pruning. As the name suggests, it is a method where we set the maximum depth to which the tree can grow.

Instability is always a concern with decision trees as a small variation in the data can result in generation of a different tree altogether.

The solution to such a scenario is ensemble learning, which we will study in the next chapter.

Decision tree learning may sometimes lead to the creation of biased trees, where some classes are dominant over others. The solution to such a scenario is balancing the dataset prior to fitting it to the decision tree algorithm.

Decision tree learning is known to be NP-complete, considering several aspects of optimality. This holds even for the basic concepts.

Usually, heuristic algorithms like greedy algorithms are used where a locally optimal decision is made at each node. This doesn't guarantee that we will have a decision tree that is globally optimal.

Learning can be hard for the concepts such as Parity, XOR, and multiplexer problems, where decision trees cannot represent them easily.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.