Unsupervised learning based patent landscapes using full-text patent data

Arho Suominen, Hannes Toivanen

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

    1 Citation (Scopus)

    Abstract

    The complexity technologies require that companies have in-depth knowledge of the nature and effect of knowledge - its depth and breadth. Companies need to master expanding technological knowledge bases creating tensions for MOT. We examine how big data in patent landscaping creates insights into MOT. Using big data to manage Competitive Technical Intelligence, companies can foster new forms of adaptive learning processes in MOT. This however requires that managers augment human judgment with machine-learning tools, prompting challenges to management traditions. We demonstrate how unsupervised learning creates insight into MOT by identifying topical knowledge foci and showing the dynamics of knowledge domains among companies. Using unsupervised learning and network analysis; we show how a semantic analysis leads to the identification of opportunities in complex environments. We illustrate this using a case in globally operating telecommunication companies using a full-text copy of USPTO-database with approximately 6 million patents data. Our results show the landscape of the companies and the underlying knowledge embedded in the companies. We discuss how managers can evaluate their technological knowledge against competitors, balancing current needs with the adoption of new knowledge. We further discuss how a semantic analysis can lead to the discovery of latent patterns and identification of opportunities.
    Original languageEnglish
    Title of host publicationManagement of Engineering and Technology (PICMET), 2015 Portland International Conference on
    PublisherIEEE Institute of Electrical and Electronic Engineers
    Pages2195-2203
    ISBN (Electronic)978-1-8908-4331-1, 978-1-8908-4332-8
    ISBN (Print)978-1-4799-1767-9
    DOIs
    Publication statusPublished - 24 Sep 2015
    MoE publication typeA4 Article in a conference publication
    EventPortland International Conference on Management of Engineering and Technology, PICMET 2015 - Portland, United States
    Duration: 2 Aug 20156 Aug 2015

    Conference

    ConferencePortland International Conference on Management of Engineering and Technology, PICMET 2015
    Abbreviated titlePICMET 2015
    CountryUnited States
    CityPortland
    Period2/08/156/08/15

    Keywords

    • big data
    • companies
    • data mining
    • industries
    • patents
    • semantics
    • unsupervised learning

    Fingerprint Dive into the research topics of 'Unsupervised learning based patent landscapes using full-text patent data'. Together they form a unique fingerprint.

  • Projects

    RAID: Radical and Incremental Innovation in Industrial Renewal

    Suominen, A. & Toivanen, H.

    1/01/1531/12/16

    Project: Business Finland project

    Cite this

    Suominen, A., & Toivanen, H. (2015). Unsupervised learning based patent landscapes using full-text patent data. In Management of Engineering and Technology (PICMET), 2015 Portland International Conference on (pp. 2195-2203). IEEE Institute of Electrical and Electronic Engineers. https://doi.org/10.1109/PICMET.2015.7273139