Revising Rules to Capture Requirements Traceability Relations: A Machine Learning Approach

In this paper we present a machine learning approach for generating requirements traceability relations. This approach is based on a new learning algorithm that produces traceability rules which are able to capture traceability relations between requirement statements specified in natural language and object models. The creation of these traceability rules is informed by examples of traceability relations which are provided by the user and is based on a generalisation of other existing traceability rules.