The Output of R&D activities: Harnessing the Power of Patents Data" 19-20 September 2013 IPTS, Sevilla IDENTIFYING EMERGING TECHNOLOGIES A Burst analysis Hélène DERNIS OECD Directorate for Science, Technology and Industry (STI) Economic Analysis and Statistics (EAS) Division
Background Joint work with : Mariagrazia Squicciarini OECD Roberto Dantas de Pinho Ministério da Ciência, Tecnologia e Inovação - MCTI Brazil
Background New technologies and new fields are emerging, also from the combination of different technologies. They take time to develop and mature. Experimentation phase (R&D or inventive activity over several years) is sometimes followed by a sudden and marked increase in innovative activity.
Burst analysis The burst (i.e. an acceleration in the development of a certain technology) can be identified with an experimental methodology that detects sharp and sustained rises in the relative frequency of patent filings in different technological areas. A burst analysis makes it also possible to identify technologies that are likely to continue booming over the next few years.
Inspired by text mining techniques Burst analysis techniques derive from Kleinberg, J. (2003), Bursty and Hierarchical Structure in Streams, Data Mining and Knowledge Discovery, Vol. 7, No. 4. http://dx.doi.org/10.1023/a:1024940629314 A fundamental problem in text data mining is to extract meaningful structure from document streams that arrive continuously over time. E-mail and news articles are two natural examples of such streams, each characterized by topics that appear, grow in intensity for a period of time, and then fade away. The published literature in a particular research field can be seen to exhibit similar phenomena over a much longer time scale. Underlying much of the text mining work in this area is the following intuitive premise - that the appearance of a topic in a document stream is signaled by a burst of activity," with certain features rising sharply in frequency as the topic emerges. ( )
DATA SOURCES
Data sources Patent applications filed under the Patent Cooperation Treaty (PCT) by filing date : 1978-2010 by IPC 4-digits codes: more than 640 subclasses. Frequency measure based on fractional counts 2 measures : 1) Burst analysis IPC occurrences in patents 2) Burst analysis IPC co-occurrences in patents e.g. combinations of IPC-4 digit pairs Threshold: More than 250 patents over the whole period
Methodology Bursts are evaluated in relative terms - patent application patterns are compared with patterns typically observed: in previous years in the technology fields considered and with what occurs in other fields. distinguish technology/field specific increases from generalised increases in overall patenting activity. Only IPC combinations with positive burst intensities are presented.
Technology burst A detected burst can be understood as a period of higher relative frequency of usage of a given IPC as compared to other periods, and normalized to the overall volume of applications in each year. Bursts are characterised by a starting and ending date and a relative burst intensity. A burst that only appears at the beginning of the reference period implies that usage of the given IPC is fading or loosing relative importance. Open ended bursts (no end year) can be used for tracking the emergence of technologies.
ACCELERATION IN THE DEVELOPMENT OF PATENTED TECHNOLOGIES IPC FREQUENCIES
Technology burst in 2000s Top 40 technologies (IPC 4-digits) by burst intensity and burst period, PCT filings, 2000-11 Figure under embargo Source: OECD, calculations based on PATSTAT, EPO, April 2013.
Technology burst in 2000s Phone and wireless communication % 1.0 0.8 0.6 0.4 0.2 0.0 Patents in selected IPC4-digits, Share in total PCT filings and burst period H04Q: SELECTING APPARATUS OR ARRANGEMENTS 155 PCT filings 2.0 % H04M: TELEPHONIC COMMUNICATION 1.5 1.0 0.5 0.0 869 PCT filings % H04W: WIRELESS COMMUNICATION NETWORKS 5.0 4.0 3.0 2.0 1.0 0.0 6433 PCT filings Source: OECD, calculations based on PATSTAT, EPO, April 2013.
Technology burst in 2000s Lighting Patents in selected IPC4-digits, Share in total PCT filings and burst period % F21K: LIGHT SOURCES 0.10 0.08 0.06 0.04 0.02 0.00 102 PCT filings % F21S: NON-PORTABLE LIGHTING DEVICES 0.25 0.20 0.15 0.10 0.05 0.00 348 PCT filings % 0.25 0.20 0.15 0.10 0.05 0.00 F21Y: INDEXING SCHEME RELATING TO FORM OF LIGHT SOURCES 284 PCT filings % 0.5 0.4 0.3 0.2 0.1 0.0 F21V: FUNCTIONAL FEATURES OR DETAILS OF LIGHTING DEVICES OR SYSTEMS THEREOF 678 PCT filings % F24J: PRODUCTION OR USE OF HEAT 0.25 0.20 0.15 0.10 0.05 0.00 359 PCT filings Source: OECD, calculations based on PATSTAT, EPO, April 2013.
ACCELERATION IN THE CO-DEVELOPMENT OF PATENTED TECHNOLOGIES FREQUENCIES IN IPC CO-OCCURENCES
Technology burst in 2000s Co-occurences Accelerations in the co-development of technologies are detected by looking at application patterns and bursts of all possible pairs of 4-digit IPC classes appearing in patent documents. Highlights the extent to which new fields arise from the cross-fertilisation of different technologies.
Technology burst in 2000s Co-occurences H05B 2006-11 H04W H04Q 1996-2001 H04N H04L H02J H01M H01L G09G G09F G07F G06F G02F G01V G01N F21Y F21V D06L C12Q C12P C12N C07K C07D C07B B60W B60L A61K Figure under embargo PCT patent applications 500 patents 1 000 patents 10 000 patents International Patent Classes (IPC) A01 - Agriculture A61 - Medical/veterinary science A63 - Sports, games, etc. B01 - Physical or chemical devices B60 - Vehicles C07 - Organic chemistry C09 - Dyes, paints, etc. C11 - Oils, fats and waxes C12 - Biochemistry C40 - Combinatorial technology D06 - Treatment of Textiles E21 - Earth/rock drilling, mining F21 - Lighting F24 - Heating, ventilating, etc. G01 - Measuring, testing G02 - Optics G03 - Photography, etc. G06 - Computing G07 - Checking-devices G09 - Educating, display, etc. G11 - Information storage H01 - Basic electric elements H02 - Electric power H04 - Electric communication H05 - Electric techniques n.e.c. A61P Source: OECD, calculations based on PATSTAT, EPO, April 2013.
Source: OECD, calculations based on PATSTAT, EPO, April 2013. Technology burst in 2000s Co-occurences H05B H04W H04Q H04N H04L H02J H01M H01L G09G G09F G07F G06F G02F G01V G01N F21Y F21V D06L C12Q C12P C12N C07K C07D C07B B60W B60L A61K Energy storage and distribution for electric and hybrid vehicles Electronic games Semiconductors Electric and hybrid vehicles Digital data acquisition, processing and transmission Figure Detergents under Chemistry and biotechnology Lighting embargo Micromolecular chemistry Displaying devices Batteries PCT patent applications Wireless and digital communication 500 patents 1 000 patents 10 000 patents 2006-11 1996-2001 International Patent Classes (IPC) A01 - Agriculture A61 - Medical/veterinary science A63 - Sports, games, etc. B01 - Physical or chemical devices B60 - Vehicles C07 - Organic chemistry C09 - Dyes, paints, etc. C11 - Oils, fats and waxes C12 - Biochemistry C40 - Combinatorial technology D06 - Treatment of Textiles E21 - Earth/rock drilling, mining F21 - Lighting F24 - Heating, ventilating, etc. G01 - Measuring, testing G02 - Optics G03 - Photography, etc. G06 - Computing G07 - Checking-devices G09 - Educating, display, etc. G11 - Information storage H01 - Basic electric elements H02 - Electric power H04 - Electric communication H05 - Electric techniques n.e.c. A61P
ON-GOING WORK
Experimental work Using data from different patenting authorities (esp. national patent offices) may underline the technology specialisation of economies and the different development trajectories of technologies across countries and over time. Tracking patent bursts over time, especially those characterised by an initially small number of patent applications, might help uncover new technological trajectories.
Experimental work Burst analysis will be included in the forthcoming OECD, Science, Technology and Industry Scoreboard, 2013 Watch out for our forthcoming OECD/STI working paper Thank you! helene.dernis@oecd.org