1
$\begingroup$

I was working on a project and was trying to validate my decisions. I wondered why would I want to use a decision tree over more powerful algorithms like random forest or Gradient boosting machine which uses similar tree based architecture.

$\endgroup$
1

3 Answers 3

1
$\begingroup$

Decision Tree has 2 key benefits over Random Forest:

  • Interpretability – In many domain, this is a must.

  • Faster prediction time – Can be an issue with RF (if n_estomators is too high) and when it has to be an online prediction.

Unless you have these constraints, RF is the go-to algorithm. For sparse and high-dimension data, SVM should also be tried.

$\endgroup$
1
$\begingroup$

Advantages:-

1.) Faster than RF.

2.) Relatively easy to interpret than other algo's, although most of the algorithms are interpretable more or less.

3.) Easy to visualize.

4.) Can have control over feature selection if you don't want to go over the filter based or wrapper based feature selection.

5.) Easy to implement.

6.) Easy to tune the hyperparameters.

7.) Easy to prune.

But if you really want to get better results, either go for RF or with some Gradient Boosting algorithms.

$\endgroup$
0
$\begingroup$

One main advantage of using Decision Tree is that you can visualize your prediction more precisely than any other classification approaches.

In addition,

  • What rules make any specific prediction? You can generate such rules

  • What feature/value is crucial for what decision? You can indirectly do the feature selection.

There are some shortcomings as well; sometimes, the tree might get quite large, so you have to reduce the size of the tree (pruning) with the cost of accuracy.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.