I'm pretty sure I did. Random Decision Forest with no regularization, no probabilistic accounting for the base rate of the target class, and with the test and

I'm pretty sure I did. Random Decision Forest with no regularization, no probabilistic accounting for the base rate of the target class, and with the test and train sets mixed rather than performing proper cross-validation. They've almost definitely overfit, and almost definitely killed someone who wasn't actually a terrorist.