Abstract
The objective of this paper is to enable organizations to generate an executable business process from high level design specifications. The basic idea is to exploit the knowledge of the existing business processes of related organizations to generate an executable business process for the given organization based on its requirements. However, this requires organizations with existing business processes to share their process execution sequences. Since the execution sequences (even after data sanitization) still include sensitive business information which organizations may not want to share with their competitors, this needs to be done in a privacy-sensitive way.
Towards this, we propose a privacy preserving approach for generating a repository of business process execution sequences. The proposed approach is based on differential privacy and does not reveal any sensitive information about individual organizations. The proposed approach is designed for a collaborative environment in which organizations share their business process details with a trusted third party called Business Process Recommendation and Composition System (BPRCS). BPRCS generates a differentially private dataset of execution sequences and employs process mining and classification techniques on this dataset to regenerate the executable business process workflow. We have implemented a prototype of BPRCS in J2EE and used it to validate the approach.