Creating test data for market surveillance systems with embedded machine learning algorithms

Market surveillance systems, used for monitoring and analysis of all transactions in the financial market, have gained importance since the latest financial crisis. Such systems are designed to detect market abuse behavior and prevent it. The latest approach to the development of such systems is to use machine learning methods that largely improve the accuracy of market abuse predictions. These intelligent market surveillance systems are based on data mining methods, which build their own dependencies between the variables. It makes the application of standard user-logic-based testing methodologies difficult. Therefore, in the context of intelligent surveillance systems, we built our own model for classifying the transactions. To test it, it is important to be able to create a set of test cases that will generate obvious and predictable output. We propose scenarios that allow to test the model more thoroughly, compared to the standard testing methods. These scenarios consist of several types of test cases which are based on the equivalence classes methodology. The division into equivalence classes is performed after the analysis of the real data used by real surveillance systems. We tested the created model and discovered how this approach allows to define its weaknesses. This paper describes our findings from using this method to test a market surveillance system that is based on machine learning techniques.