Using Markov Models and Statistics to Learn, Extract, Fuse, and Detect Patterns in Raw Data

TL;DR

This paper presents data-driven methods for extracting and utilizing Markov models from raw data to identify patterns, with applications in shipping, cybersecurity, and radiation detection.

Contribution

It introduces novel approaches for deriving stochastic state machines directly from observed data, enhancing pattern detection in various practical domains.

Findings

Effective extraction of Markov models from raw data

Successful application in shipping pattern analysis

Detection of botnet activities and radiation sources

Abstract

Many systems are partially stochastic in nature. We have derived data driven approaches for extracting stochastic state machines (Markov models) directly from observed data. This chapter provides an overview of our approach with numerous practical applications. We have used this approach for inferring shipping patterns, exploiting computer system side-channel information, and detecting botnet activities. For contrast, we include a related data-driven statistical inferencing approach that detects and localizes radiation sources.