Applying artificial intelligence and machine learning techniques in the internal audit environment: A simulation test of an improved algorithm for detection of fraudulent events in financial entity transaction processing operations

This paper describes a new approach to financial fraud detection that is informed by both artificial intelligence and machine learning. It proposes, then tests with a simulated database of transactions, a metaheuristic algorithm based on two-layer deep learning. Employing a simple adaptation to a basic deep learning model that applies a hybridized honey badger modification, demonstrably higher quality outcomes were obtained than with commonly used competing models.