Intelligent Extraction of a Knowledge Ontology From Global Patents: The Case of Smart Retailing Technology Mining

• Published in: International journal on semantic web and information systems Vol. 16; no. 4; pp. 61 - 80
• Main Authors: Trappey, Charles V, Chang, Ai-Che, Trappey, Amy J. C
• Format: Journal Article
• Language: English
• Published: Hershey IGI Global 01.10.2020
• Subjects: Algorithms; Bibliographic coupling; Business competition; Clustering; Data mining; Dirichlet problem; Forecasts and trends; Information systems; Innovations; Intellectual property; Inventors; Knowledge; Knowledge representation; Mineral industry; Mining industry; Ontology; Probability; Retail industry; Retailing; Semantic analysis; Semantic web; Semantics; Trends; User experience
• ISSN: 1552-6283; 1552-6291
• DOI: 10.4018/IJSWIS.2020100104

The growth of global patents increased over the last decade as enterprises and inventors sought greater protection of their intellectual property (IP) rights. Global patents represent state-of-the-art knowledge for given domains. This research develops a hierarchical Latent Dirichlet Allocation (LDA)-based approach as a computational intelligent method to discover topics and form a top-down ontology, a semantic schema, representing the collective patent knowledge. To validate the knowledge extraction, 1,546 smart retailing patents collected from the Derwent Innovation platform from 2011 and 2016 are used to build the domain ontology schema. The patent set focuses on in-use, globally established, and non-disputed IP covering payment, user experience, and information integration for smart retailing. The clustering and LDA-based ontology system automatically build the knowledge map, which identifies the technology trends and the technology gaps enabling the development of competitive R&D and management strategies.