A Machine Learning Attack on Illegal Trading

We design an adaptive framework for the detection of illegal trading behavior. Its keycomponent is an extension of a pattern recognition tool, originating from the field of signalprocessing and adapted to modern electronic systems of securities trading. The new methodcombines the flexibility of dynamic time warping with contemporary approaches from extremevalue theory to explore large-scale transaction data and accurately identify illegal trading patterns.Importantly, our method does not need access to any confirmed illegal transactions fortraining. We use a high-frequency order book dataset provided by an international investment firm to show that the method achieves remarkable improvements over alternative approaches inthe identification of suspected illegal insider trading cases