Mining and Matching Relationships From Interaction Contexts in a Social Manufacturing Paradigm

There is an increasing use of social interaction contexts in the cross-enterprise manufacturing problem solving. To transform these massive and unstructured data into decision-support information for cross-enterprise manufacturing demand-capability matching, we present automated solutions to two phases: 1) extracting relationships based on a semi-supervised learning approach to derive formalized heterogeneous manufacturing network from the unstructured text-based context that contains high levels of noise and irrelevant information and 2) matching group-level relationships among the entities in the established manufacturing network. The extracting phase formulates network data using multiattributed graph that can encode various entities and relationships. The matching phase is based on probabilistic multiattributed graph matching, and implemented using distributed message passing algorithm. We developed a prototype system to verify the proposed model, which is also flexible to new domains of contexts and scale to large datasets. The ultimate goal of this paper is to facilitate knowledge transferring and sharing in the context of cross-enterprise social interaction, thereby supporting the integration of the resources and capabilities among different enterprise.