Outlier detection in the concept drifting environment

Abstract

Outliers are observations that lie far away from the fitting function deduced from the bulk of a set of observations. The outlier detection has become more challenging when the nature of data has involved with the `concept drifting.` To address this challenging issue, this study explores a decision support mechanism (DSM) for coping with the outlier detection problem in the concept drifting environment. The proposed DSM has the following features: (1) implementation of the resistant learning with envelope module via the adaptive single layer feed-forward neural network, (2) implementation of the incremental learning concept via the moving window technique, and (3) effectiveness and efficiency in terms of being more accurate in identifying outliers and of having to further investigate fewer outlier candidates. An experiment is implemented to validate the proposed DSM and the results are promising.