We have proposed replicator neural networks (RNNs) as an outlier detecting algorithm [15]. Here we compare RNN for outlier detection with three other methods using both publicly available statistical datasets (generally small) and data mining datasets (generally much larger and generally real data). The smaller datasets provide insights into the relative strengths and weaknesses of RNNs against the compared methods. The larger datasets particularly test scalability and practicality of application. This paper also develops a methodology for comparing outlier detectors and provides performance benchmarks against which new outlier detection methods can be assessed.
Graham J. Williams, Rohan A. Baxter, Hongxing He,