Accounting errors are quite commonplace in businesses and government organizations. These can include errors occurring due to improper record keeping and data entry, missing data, irregular checks and balances or errors made in principle. These errors are sometimes easily rectifiable -- for example, by imputation of missing values and other standard statistical measures; but at other times, they can be costly leading to major financial penalties from non-compliance for the business or organization in question.
Our goal in this research is to identify plausible accounting errors using sophisticated artificial intelligence / machine learning powered techniques. The identification and detection of such errors very early in the process can help prevent regulatory and other missteps down the road and are therefore very useful for the businesses in question.