CollabXplore: Mining and Recommendation of Interdisciplinary Academic Collaboration

The project introduces CollabXplorer, a system designed to facilitate interdisciplinary collaborations and the discovery of potential collaborators. Global semantic information and social influence are utilized to create a scholar embedding model, which converts academic information into vector representations to identify potential collaborators from various scholarly fields. To suggest suitable collaborators, the system applies a recommendation algorithm based on user prompts and scholars’ influence. User-friendly visual interface aids in presenting the system’s outcomes. We also conducted evaluation and testing experiments, which show that the system provides a good scholar vector representation and high recommendation accuracy, thus, making it suitable for large-scale applications.