An Introduction to Outlier Ensembles

Abstract

The outlier analysis problem has been widely studied by database, data mining, machine learning and statistical communities. Numerous algorithms have been proposed for this problem in recent years (Aggarwal, Outlier Detection in High Dimensional Data, [6]; Angiulli, Fast Outlier Detection in High Dimensional Spaces, [9]; Bay, Mining distance-based outliers in near linear time with randomization and a simple pruning rule, [11]; Breunig, LOF: Identifying Density-based Local Outliers, [14]; Knorr, Algorithms for Mining Distance-based Outliers in Large Datasets, [35]; Knorr, Finding Intensional Knowledge of Distance-Based Outliers, [36]; Jin, Mining top-n local outliers in large databases, [39]; Johnson, Fast computation of 2-dimensional depth contours, [40]; Papadimitriou, LOCI: Fast outlier detection using the local correlation integral, [53]; Ramaswamy, Efficient Algorithms for Mining Outliers from Large Data Sets, [55]).