Using non-supervised machine learning to generate a knowledge ontology for patent analytics

Amy J.C. Trappey*, Charles V. Trappey, C. H. Zou

*Corresponding author for this work

Research output: Contribution to journalConference article

Abstract

The growth of patenting activities has increased as a result of enterprises seeking greater intellectual property protection and commercialization using global patent systems and legal conventions for sales or licensing. This research develops a novel system to intelligently generate the knowledge ontology for patent analytics. The intelligent system automatically tracks patent topics and searches related patents across multiple patent databases. The first step is to cluster the domain patents into key groups without supervision. Latent Dirichlet Allocation (LDA), an unsupervised machine learning approach, is modified to identify key topics and to match patents with similar topic categories. The system extracts key terms and constructs the ontology. The technical and functional terms are depicted using links across ontology nodes that enable enterprises to target research and development investments that best support their patent portfolios and market investment strategies.

Original languageEnglish
JournalProceedings of International Conference on Computers and Industrial Engineering, CIE
Volume2018-December
StatePublished - 1 Jan 2018
Event48th International Conference on Computers and Industrial Engineering, CIE 2018 - Auckland, New Zealand
Duration: 2 Dec 20185 Dec 2018

Keywords

  • Latent dirichlet allocation
  • Ontology
  • Patent clustering
  • Technology mining

Fingerprint Dive into the research topics of 'Using non-supervised machine learning to generate a knowledge ontology for patent analytics'. Together they form a unique fingerprint.

  • Cite this