DECISION SUPPORT MODEL FOR INTELLECTUAL PROPERTY MANAGEMENT

Abstract

This article examines the pressing issue of improving the effectiveness of decision support in intellectual property management. A well-founded definition of the term  decision support  is provided in the context of intellectual property management using artificial intelligence technologies. An original decision support model for managing intellectual property assets in government agencies is presented, taking into account the basic stages of intellectual property management and the requirements specific to public administration. The developed model formalizes the procedure for selecting and combining intelligent
algorithms from a variety of those acceptable for specific business conditions, thereby improving the validity of decisions, including in the presence of incomplete data. The methodological foundation for the
developed algorithms includes graph neural networks with a knowledge graph-based model and the GAT algorithm, genetic programming and particle swarm methods, a gradient boosting algorithm combined with BERT contextual embeddings, and the K-means method with contrastive loss.