A novel sustainable value co-creation improvement framework based on group attractiveness matching-graph neural network

Purpose Sustainable value co-creation has become a critical target in open innovation communities (OICs). This study aims to provide a novel sustainable value co-creation improvement framework by promoting high-quality innovation resources. Design/methodology/approach This study constructs a novel sustainable value co-creation improvement framework based on the group attractiveness matching-graph neural network (GAM-GNN). The framework describes the OIC as a user-idea hypernetwork structure and comprises three modules: group attractiveness characteristics monitoring, diverse-orientation matching and effective interaction recommendation module. Findings The experiment uses data from a well-known smartphone community. Our method outperforms existing models. The accuracy is 84.63, and the F1 score is 0.8036. The accuracy of our method increases by 2.96% for GraphSage, 3.72% for graph isomorphism network, 20.40% for graph-auto encoder, 51.88% for graph convolutional network, 9.14% for graph attention network and 66.92% for convolutional neural network. Originality/value This study extends the effectiveness perspective of sustainable value co-creation by integrating insights from group dynamics theory and group attractiveness effects. This provides practical guidance for enterprises to improve sustainable value co-creation.