Due to the way that decision trees work, do random forest classifiers always get 100% accuracy on their own training data? My random forest classifier got 100% accuracy on its own training data, so I'm not sure if this is a fluke or if there is a logical reason for this. The model got 99.4% accuracy on its testing data, so it may have just had high accuracy in general.
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
0
No, random forest is not guaranteed to get 100% accuracy on the test data. But it also doesn’t mean it’s overfit because relatively high scores on test data are common with random forests.
The trees of a random forest are trained to leaf purity. BUT, they are only trained on a subset of the samples/features. A pure decision tree rarely generalizes well so samples not in a given tree's subset of the training samples might be wrong. The voting mechanism of random forest tends to even out all of the mistakes and provide a good score but a sample in a training set may be so misrepresented by enough decision trees to outweigh the correct votes resulting in an incorrect prediction.
