Enabling Business Rule Traceability as Guidance to Regulation Origination in Finance

Abstract

We evaluate and assess the applicability of a Reference Architecture for Traceability (RA4T) to Finance regulations, and assess if it can assist with traceable implementation of regulations. This enables tracing back business rules to their source in regulations. This evaluation is conducted in a thesis circle in which other members assess the architecture in other domains. This reference architecture aims to guide developers in making rule-based systems which provide traceability. Traceability ensures that end users in their interface can trace back the full origin of system conclusions to their policy documents.
The reference architecture applies a Controlled Natural Language (CNL) to rule-development to ensure consistent interaction between humans, processable by computers. This research addresses two areas: First, we will evaluate the effects on traceability by applying RA4T to Finance Regulations and see whether traceability changes for end users. The context is European Anti Money Laundering Directives, as they stipulate traceable implementation. Then, we will structure the scenarios in CNL and transform to code. Logical conclusions on rules are tied to data to accommodate basic explanations by providing snippets of the source regulation. The challenge is keeping intact source-related meta-data needed for traceability. This contribution is sufficient when the Architecture is evaluated via scenarios, and constitutes an improvement in the domain context.

Date of Award	4 Jul 2022
Original language	English
Supervisor	Lloyd Rutledge (Examiner) & Ella Roubtsova (Co-assessor)