nBoost is a suite of boosting algorithms designed to solve binary classification problems on data that is not linearly separable by a convex combination of base hypotheses, i.e. noisy data. WARNING: Active development. Underlying algorithm is unstable.