Bisociation networks analysis for business process

Bisociation Network (BisoNet) is a novel approach for creative information discovery, and it can be projected to many real application domains. Bisociation of business processes onto a network is one of such applications. In this paper, we investigate business processes on the BisoNet, and develop a directed graph model to map the relations between business process flows. Based on the BisoNet model, we analyze the real-world data provided by a call service centre. The network-based statistics show that some special process steps could be key steps that greatly affect the performance of the service, and could result in a case unsolved. The network is simplified through constructing the network with shortest path of each process flow, and the simplified network may represent an optimal process pattern. This may provide a reference to the business organization for improving the quality of their service.