Applying Rule Learner to IRS Historical Data

In 2007 OpenRules applied its Rule Learner, a machine-learning component of its popular open-source business rules management system, to the Internal Revenue Service (IRS) National Research Program data provided by NHQ Research.  As a result, the Rule Learner was able to successfully generate business rules that model the issue selection domain.  OpenRules applied different ML algorithms to a large volume of historical data and generated the learned rules in a form automatically executable by the OpenRules engine.  The achieved classification results are comparable to or better than results previously achieved by manual tax return selection.   The resulting approach demonstrated a generic, yet fully automated integration of the learned rules into a Business Rules Management System.  Read the Press Release.

In 2008 IRS awarded OpenRules, Inc. another contract, “Integrating Machine Learning Models With Business Rules Environment”, with the objective of integrating machine learning results into a standard business rules environment in a way that allows IRS subject matter experts to easily identify any gaps between the rules developed by its experts to detect potential non-compliance, and the rules automatically developed by the machine learning system. Read the Press Release.

During 2009-2012 OpenRules successfully completed two more IRS contracts related to the integrated use of machine learning and business rules technologies: “Integrating Machine Learning Models With Business Rules Environment” and “Machine Learning Models Research and Development with NRP and LMSB Data”. 

The IRS Performance Report stated: “IRS Management was satisfied with the services being performed by the Contractor and its staff members. No complaints about the key personnel working under this order. The overall performance of Dr. Jacob Feldman and staff on behalf of The OpenRules, Inc., was very good. This company would be highly recommended to work for the Internal Revenue Service in the future in regards to Machine Learning (ML) Models Performance in IRS Enforcement Programs.”