Suitability of Random forest methods for n < p classification problem?



Is Randomforest a good approach for (n less than p) classification problems ?

here n is the number of observations(rows)
p= number of features(variables)
and n is less than p situation



Random Forest typically would not perform well in these kind of problems. Actually, the performance starts degrading even when n & p become comparable.

I would look at ways to reduce the number of features first before applying Random Forest.



Thanks will do some variable reduction like Pricipal component analysis or some regularisation