Trust and Risk Relationship Analysis on a Workflow Basis: A Use Case

Trust and risk are often seen in proportion to each other; as such, high trust may induce low risk and vice versa. However, recent research argues that trust and risk relationship is implicit rather than proportional. Considering that trust and risk are implicit, this paper proposes for the first time a novel approach to view trust and risk on a basis of a W3C PROV provenance data model applied in a healthcare domain. We argue that high trust in healthcare domain can be placed in data despite of its high risk, and low trust data can have low risk depending on data quality attributes and its provenance. This is demonstrated by our trust and risk models applied to the BII case study data. The proposed theoretical approach first calculates risk values at each workflow step considering PROV concepts and second, aggregates the final risk score for the whole provenance chain. Different from risk model, trust of a workflow is derived by applying DS/AHP method. The results prove our assumption that trust and risk relationship is implicit.