Autoencoders

Two deliberately lopsided models — one that never calls a positive wrongly, one that never misses a positive — and a meta-model that reads both.

Python · XGBoost · TensorFlow · Ensemble methods