ABSTRACT

This chapter presents approaches that use length statistics of emerging patterns (EPs) to detect outliers, and use EPs in rare-class classification.

An outlier is “an observation which deviates so much from other observations as to arouse suspicion that it was generated by a different mechanism”

[184]. Outlier detection is a core task in data mining, and has attracted a great deal of attention from the research communities [189, 2, 215, 231].