Statistical Interpretation of Technologies Using IPC Codes in Patents

• Activity Number: 137
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Daiho Uhm*+ and Sunghae Jun
• Companies: University of Arkansas at Fort Smith and Cheongju University
• Keywords: Patent ; Patent analysis ; IPC ; Factor analysis ; Clustering

Abstract

A patent is an exclusive right of intellectual properties to an inventor or its assignee during a limited time period. Patent analysis suggests objective and accuracy methods to analyze patent data and forecast technologies in future. In this paper some statistical analyzing tools are applied to the number of international patent classification (IPC) codes. The IPC, established by the Strasbourg Agreement 1971, is a hierarchical system for the classification of patents based on the technology areas. Taking the first four digits of the IPC codes, the numbers of the IPC codes in each patent document are used to provide the relationship among the technology managements. We apply statistical data analysis such as correlation, factor analysis and clustering to a patent data set in which the patents were registered by Hyundai motor company by July, 2013. The relationships among the given IPC codes are described in terms of a few underlying random quantities, and the patent documents are clustered by IPC system. Our research might contribute meaningful results to R & D planning and technology management.