Machine learning strategies are typically divided into 3 wide categories, which correspond to learning paradigms, dependant upon the nature from the "signal" or "responses" accessible to the learning method: A few wide categories of anomaly detection techniques