Computational Statistics for Cybersecurity

Abstract

This article discusses several applications of computational statistics for cybersecurity. Nonparametric probability density estimation is widely used in computational statistics, and has many applications in cybersecurity, both for understanding and analyzing data and for classification and anomaly detection. These methods are described and applied to network flows and malware characterization. Streaming data methods are discussed and illustrated on the network flows data. Methods for analyzing graphs and time series of graphs are discussed and illustrated.