Traceability links connect artifacts in software engineering models to allow tracing in a variety of use cases. Common to any of these use cases is that one can easily find related artifacts by following these links. Traceability links can significantly reduce the risk and cost of change in a software development project. However, finding, creating and maintaining these links is costly in most cases. In any real-world project of significant size the creation and maintenance of traceability links requires tool support.In this paper, we propose a novel approach to support the automation of traceability link recovery based on Association Rule Mining and operation-based change tracking.Traceability link recovery is the activity of finding missing or lost traceability links. Our approach automatically generates a list of candidate links based on the project history along with a score of support and confidence for every candidate link. We transformed the data from an operation based change tracking system to sets of frequent items,which serve as input for Association Rule Mining (ARM).We applied our approach to data from a software development project with more than 40 developers and assessed the quality of the candidate links in interviews.