Binary Information Press
Outlier detection is a prominent research domain in data mining. Numerous approaches have been proposed to discover anomaly objects in data populations. Detecting outlying subspaces is a relatively new research problem in outlier detection domain. Most existing studies of outliers focus on detecting them, yet limit attention has been paid to the problem of finding the intentional knowledge of the outliers. This paper is concerned with the problem of discovering sets of attributes that account for the abnormality of outliers belong to a class within a given dataset.