Automated Conflict Resolution in CDSS using Community Feedbacks

In collaborative data sharing systems, groups of users usually work on disparate schemas and database instances, and agree to share the related data among them (periodically). Each group can extend, curate, and revise its own database instance in a disconnected mode. At some later point, the group can publish its updates to other groups and get updates of other ones (if any). The reconciliation operation in the CDSS engine is responsible for propagating updates and handling any data disagreements between the different groups. If a conflict is found, any involved updates are rejected temporally and marked as deferred. Deferred updates are not accepted by the reconciliation operation until a user resolves the conflict manually. In this paper, we propose an automated conflict resolution approach that depends on community feedbacks, to handle the conflicts that may arise in collaborative data sharing communities, with potentially disparate schemas and data instances. The experiment results show that extending the CDSS by our proposed approach can resolve such conflicts in an accurate and efficient manner.