Large-scale cyber attacks monitoring using Evolving Cauchy Possibilistic Clustering.

•A novel evolving possibilistic Cauchy clustering (eCauchy) is presented that is able to learn a classifier in an online manner on a stream of data.•As oppose to some evolving algorithms the presented approach has only few tuning parameters.•The eCauchy clustering is tested on large-scale monitoring for cyber-attacks on a KDD data set.•The results are given for all three KDD data sets in a form of typical classifier goodness measures.•The obtained results are promising and show that the approach can be potentially useful for monitoring network traffic.