A New Language and Input–Output Hidden Markov Model for Automated Audit Inquiry
Pushkin P.E. Kachroo, Aaron Saiewitz, Robyn L. Raschke, Shaurya Agarwal, Archie J. Huang · IEEE Intelligent Systems · 2020
This article presents a mathematical coding language to express dynamic interactions between auditors and client personnel. Then, an input-output hidden Markov model is presented that represents clients as well as auditors, and models the coupled system. The calibrated model can be used to design optimal automated auditors, and can also be used to perform analysis of client inquiry responses. A case study is performed using data collected with subjects simulating auditor-client communications in a controlled environment. We also discuss the details of model calibration, validations, and significance of results.