Mastering Java Machine Learning

Outlier or anomaly detection

Grubbs, in 1969, offers the definition, "An outlying observation, or outlier, is one that appears to deviate markedly from other members of the sample in which it occurs".

Hawkins, in 1980, defined outliers or anomaly as "an observation which deviates so much from other observations as to arouse suspicions that it was generated by a different mechanism".

Barnett and Lewis, 1994, defined it as "an observation (or subset of observations) which appears to be inconsistent with the remainder of that set of data".

Outlier algorithms

Outlier detection techniques are classified based on different approaches to what it means to be an outlier. Each approach defines outliers in terms of some property that sets apart some objects ...