Classification error and training time of Random Forest and Randomer Forests variants for all 121 datasets from the XXX benchmark comparison paper \cite{???}. (A) Average (dots) and 0.1 standard deviation level sets (dashed lines) for each method. (B) Classification error of Randomer Forest (sparse+delta+robust) vs. that of Random Forest for each of the 121 datasets. The black line indicates equal classification error of the two algorithms. Color indicates dimensionality of the datasets and size of points indicates number of samples. Note that RerF almost always does as well, and often significantly better.
Classification error and training time of Random Forest and Randomer Forests variants for all 121 datasets from the XXX benchmark comparison paper \cite{???}. (A) Average (dots) and 0.1 standard deviation level sets (dashed lines) for each method. (B) Classification error of Randomer Forest (sparse+delta+robust) vs. that of Random Forest for each of the 121 datasets. The black line indicates equal classification error of the two algorithms. Color indicates dimensionality of the datasets and size of points indicates number of samples. Note that RerF almost always does as well, and often significantly better.