In the following link the answer recommends a feauture amount of N/3 for regression (or it is quoted).
Where N corresponds to the sample size:
How many features to sample using Random Forests
Is there any paper which quotes this?
Asked
Active
Viewed 45 times
1 Answers
0
Not sure what is meant by a paper. Are you asking if there is mathematical proof that this is the best setting all of the time?
There are some experiments pointed to from here, a text book quote in that answer, and in the link you posted.
The answer is "it depends". You can tune this parameter for your data and problem. Perhaps n/3 is a good place to start and maybe close enough to the optimum that it does not need to be further tuned, compared to the other parameters and the time you have.
Craig
- 1,034
- 5
- 10