Semi-automatic service value network modeling approach based on external public data

Various emerging IT technologies are widely used in the service industry. Thus, an increasing number of new service models have also emerged, including the Internet of Services (IoS). The IoS supports network-based service collaboration and transactions among various service participants from different domains and different organizations, and it is expected to deliver the maximum service value to all stakeholders. To describe the cross-domain, cross-organization, and cross-value chain characteristics of the IoS from a value perspective and support subsequent analysis of the value network and optimization of the IoS, this paper proposes a semi-automatic modeling method for a IoS-oriented value network based on external public data. We first propose an intelligent domain entity recognition algorithm based on multidimensional web data to help value network modelers realize effective and efficient recognition of service participants. Then, based on external news data, an intelligent domain relationship extraction algorithm that combines the Bert + BiLSTM + CRF model with the LightGBM model is proposed to effectively and efficiently identify the value exchange relationships among service participants, thereby forming an IoS-oriented value network model (IVN). Finally, to extend the cross-domain semantics of the IVN and support analysis of the IVN, we present a domain-specific value chain extraction algorithm based on typical patterns to complete the cross-domain semantic annotation of the IVN. The effectiveness and efficiency of the proposed methods and algorithms are validated through experimental analysis and a case study, which can be of great help in IVN modeling.