Rule induction from exceptions and outliers for continuous auditing

Abstract

This manuscript proposes using a feedback loop built in a form of rule induction from confirmed instances of misstatements identified through exception and outlier analyses to update existing rules or create new ones for additional filter discovery. This addition to the analytics of the continuous audit informs auditors about potential misstatement risks they might not be aware of and suggests updates and new filters that could be integrated into the transaction verification module. The proposed artifact enhances the use of risk scores in transaction verification by better reflecting the underlying causes of misstatements. Analyzing confirmed outlier instances with rule induction algorithms enables auditors to develop additional helpful rules to identify more exceptions—rules they may not be aware of. I believe that these extensions could become an important part of the continuous audit framework.