I'm trying to conduct a research using decision trees. Is it possible to build a random forest for each algorithm? (ex. a random forest using the ID3 algorithm)?
And would it be possible to create a mixed random forest? (ex. a random forest using an ID3 and C4.5 decision tree model)
I'm planning to use Python in doing this.
Hello Leejames Merlas
A Random Forest is just an ensemble of trees, depending on the implementations the type of decission trees may vary. You could use the scikit-learn
sklearn.ensemble.BaggingClassifier module to bag some tress. For example:
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import BaggingClassifier
tree = DecisionTreeClassifier()
bag = BaggingClassifier(tree, n_estimators=100, max_samples=0.8,
random_state=1)
bag.fit(X, y)
You may see an implementation of it in: https://jakevdp.github.io/PythonDataScienceHandbook/05.08-random-forests.html
Best regards,
Unai
ID3 builds a decision tree maximizing information gain at every split recursively. Thus it uses all variables in your dataset unless told otherwise.
I'd also look into the R package ranger which implements random forests.
Greetings from Germany Philipp
Philipp Packmohr thanks for the info but from your answer, is it correct to assume that it is possible to do what I mentioned above then?
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python