We presented a large-scale benchmark experiment for comparing the performance of logistic regression and random forest in binary classification settings. The overall results on our collection of 243 datasets showed better accuracy for random forest than for logistic regression for 69.0% of the datasets. On the whole, our results support the increasing use of RF with default parameter values as a standard method—which of course neither means that it performs better on all datasets nor that other parameter values/variants than the default are useless!
