Technology analysis using patent citation network of a seminal patent final

SECOND INTERNATIONAL CONFERENCE ON BUSINESS ANALYTICS AND
INTELLIGENCE (ICBAI-2014), BANGALORE
Technology analysis using patent citation network of a seminal patent
Shanmukha Sreenivas P
Research Engineer
INDUS TechInnovations
Bangalore 560071, INDIA.
E-mail:shanmukha@industechinnovations.com
Abstract:
There exists a need for a quick technology intelligence process to visualize, identify and track changes in
science and technology for technology driven entities. In order to fill this gap, a methodology to portray a
patent citation network of a seminal patent is proposed here. This methodology can be primarily used to
determine what sub-technologies contributed to the technology in question, at what period in time and
with how much impact. The same can also be extended to identify what sub technologies spun out from
the technology/invention and also the application areas of the same. Patent latent variables like strength
and similarity are also used to illustrate the development and flow of scientific knowledge in the
technology considered. The efficacy of the methodology is demonstrated by considering patents related to
light emitting GaN based compound semi-conductor devices technology (Blue LED Technology), and
analyzing the findings resulting from the use of the methodology.
Keywords: Patent Citation Network; Technology Trajectories; Blue LED Technology; Graph layout
Introduction:
The analysis of science and technology for R&D focus, strategic capacity building and for policy making
is a tedious, multi-fold and a complicated process. There is a need for a QTIP: Quick Technology
Intelligence Process to visualize, identify and track changes in science and technology for technology-
driven businesses (Porter, 2005).
This research paper responds to the above need by presenting a methodology to portray a patent network
map that represents citations of patents pertaining to a technology of interest among other patent
variables. This methodology can be primarily used to determine what sub-technologies contributed to the
technology in question, at what period in time and with how much impact. The same can also be extended
to identify what sub technologies spun out from the technology/invention and also the application areas of
the same. At a higher level, it signifies the development and flow of scientific knowledge in a particular
field. Jaffe et al. conducted pioneering research on the construction and analysis of a knowledge spillover
network using patent citation data (Jaffe et al., 1993). The use of patent citations to provide a trace of
knowledge transfer (Jaffe et al., 1998), to detect international technology diffusion (Jaffe et al., 1999) and
in analyzing competitive advantages to firms (Patel D. et al., 2011) has been widely researched. The
patent citation layout proposed here, can also be also used to draw insights on technology proximity and
licensing opportunities related to the technology being analyzed. The potential of the methodology is
demonstrated by considering patents related to light emitting GaN based compound semi-conductor
devices technology, and analyzing the findings resulting from the use of the methodology.

Methodology:
A seminal patent related to the technology of interest was identified and the forward and backward patent
citations were extracted. Multiple levels of forward and backward citations of the resulting set can be
considered based upon computational limitations. Various attributes of the resulting patent set were
extracted from the patent database using Relecura, a patent and portfolio analysis platform and further
processing was done to tag each patent to a particular sub-technology. In order to render a patent citation
network, a table consisting of all the nodes (patents) along with attributes and a table indicating directed
edges was created. This graph was rendered with the patents pegged in a chronological order along
longitudes and also arranged based on technology proximity along latitudes using Gephi, an open source
software for exploring and manipulating networks.
The above patent citation network methodology was applied to study “Light emitting GaN based
compound semi-conductor devices Technology”. Work on light emitting GaN compound based
semiconductors was started way back in the early 70s by Maruska followed by Pankove and Miller
(Maruska, 1990). Research on developing defect free GaN crystals was also started by Isamu Akasaki and
Hiroshi Amano at the Nagoya University in Japan. However, it was Shuji Nakamura of Nichia
Corporation who laid the keystone in this field of research by producing a visible light emitting, high
efficiency, double-hetero structure semiconductor device. The analysis is based upon this particular
invention and hence the patent, US5578839 was treated as the seminal patent (Holloway et al., 1995).
In order to understand what technologies contributed to make this happen and what technologies evolved
downstream, making use of this invention, the citation network of this patent is analyzed, extending to
two levels. The forward citations and their forwards of the seminal patent, and the backward citations and
their backwards of the seminal patent are considered for this analysis.
The ForceAtlas2 and the Geo Spatial Layout graph rendering algorithms were used to understand
knowledge transfer patterns and portray the landscape in a comprehensible manner.
The ForceAtlas2 algorithm is a force-directed layout developed by the Gephi team (Jacomy et al., 2014).
In any force-directed layout, spring-like attractive forces based on Hooke's law (Fa= -k.d) are used to
attract pairs of endpoints of the graph's edges towards each other, while simultaneously repulsive forces
like those of electrically charged particles based on Coulomb's law (Fr= k/d2
) are used to separate all pairs
of nodes. A detailed description on various force-directed graphs and variations of the forces used are
described in detail by Kobourov, 2013.
Further, the structure of relationships between patents and the degree of similarity was better understood
by tweaking the layout algorithm used and also by using user defined functions to populate graph
parameters viz. node size, node color, edge thickness, edge color among many others. The resulting
technology layout was also rendered for various players in that particular technology domain, providing
insights on competitive intelligence and knowledge spillovers.
For this study, Relecura was used to obtain patent metrics, citations and technology categories and Gephi
was used to render the graphs.
Hall, Jaffe and Trajtenberg analyzed a set of more than four million US patents and have bucketed US
and IPC classes into 36 technological buckets using Factor Analysis which is widely used by researchers
to portray patent overlay maps and to identify tech clusters and patent thickets. This dataset is made
available to public on the NBER (National Bureau of Economic Research, US) website. These categories
are used in this analysis to identify tech clusters in the citation network.
Going a step further, Relecura Technology Categories were also used in this analysis to obtain a more
granular understanding of the technology landscape. Defining and portraying the importance of a patent

by merely using the number of forward citations accumulated doesn’t provide an accurate picture, hence
the Relecura Patent Quality metric was also used to indicate the value of a patent in the citation network.
The Relecura Patent Quality is a proprietary metric related to patent valuation, quality and monetization
potential of each individual patent.
Results:
The seminal patent US5578839 (Multilayer elements with indium gallium nitride on clad layers, dopes
for p-n junctions) was selected in the area of “Light emitting GaN based compound semi-conductor
devices Technology” to demonstrate the use of patent citation networks in performing a technology
analysis.
Table 1. The seminal patent considered for the patent citation network analysis
Publication
Number
Title Filing Date Inventors Original
Assignees
US5578839A Light-emitting gallium
nitride-based compound
semiconductor device
1993-11-17 NAKAMURA
SHUJI|MUKAI
TAKASHI|IWASA
NARUHITO
NICHIA
CORP
For the seminal patent in table 1, two level forward and backward citations were extracted.
Total patents (Nodes): 2979
Total connections (Edges): 5302
In order to tag each patent to a particular sub-technology, as this helps in visualizing tech clusters; the
first IPC, CPC, US classifications and the Relecura Technology categories were used for each patent.
The first IPC/CPC class mentioned on a patent may not necessarily reflect the patent’s core sub-
technology. However, the first US classification should be more closely related to the patent’s core
technology area, as it is defined as the main/primary class by the examiner. As relying only on the US
primary class would limit the dataset only to US patents both the classifications are used in this analysis.
Another means of circumventing this problem is by employing Relecura Technology categories as these
categories are based on curated information.

Figure 1. The two level patent citation network of the seminal BLUE LED patent using the ForceAtlas2
algorithm.
Figure 1 depicts the patent citation network rendered using the ForceAtlas2 algorithm in Gephi. Highly
connected nodes are attracted and loosely connected nodes are repelled at the same time. The connections
refer to citations and the nodes to patents. The patent citation network is a directed graph, curved edges
are used to indicate direction of citation. The direction of knowledge flow can be understood by following
the curves in a clockwise direction. The size of the node indicates the number of forward citations a
patent has accumulated, thus indicating its importance.
The following observations can be made from the two level patent citation network portrayed in figure 1.
1) A majority of the network nodes belong to the Semiconductor category with these patents densely
co-citing each other.
2) Drugs & Medical Chemistry category (bright turquoise) patents are also observed as a distinct
cluster at the bottom left indicating that GaN based LEDs have also had a prominent use in
medical applications. A few of these patents are related to the use of blue LEDs in dental curing
apparatuses.

The photolithography (dark blue), Lighting (purple) and optics (magenta) categories are spread across the
entire network. Indicating that these sub-technologies are very closely associated in developing this
technology.
The same layout is rendered with the NBER categories replaced with the Relecura Technology categories
and the Count of forward citations replaced with the Relecura Patent Quality metric in Figure 2.
Figure 2. The two level patent citation network of the seminal patent using Relecura Technology
categories and Relecura Patent Quality.
Figure 2 illustrates the two level patent citation network of the seminal patent rendered with ForceAtlas2
algorithm with the color codes indicating Relecura Technology categories and the size of the node
proportional to the metric Relecura Patent Quality. The Relecura Patent Quality metric is discrete and
varies from 0.5 to 5 with increments of 0.5. It can observed form the graph that many of the highly rated
patents pertain to the categories ‘Electric elements – Semiconductor devices’ and ‘Electric Heating and
Lighting’.

However, on using the ForceAtlas2 algorithm to render the patent citation layout, it is not clear as to what
the areas are from which the seminal invention was built upon and the areas of application of the
invention.
In order to obtain greater clarity regarding the technologies that contributed to the Blue LED invention
and the technologies that panned out of this invention, the same graph is plotted with the nodes (patents)
pegged in a chronological manner.
In this Geo-Spatial Layout, the Longitudes correspond to the year of filing of the patent, and the latitudes
correspond to different technology buckets.
Figure 3. The patent citation network of the seminal patent in a Geo-Spatial mode using IPC based NBER
categories
Figure 3 illustrates the two level patent citation network in the Geo-Spatial mode with the size of the node
indicating the number of forward citations the corresponding patent has accumulated and the color codes
representing various IPC based NBER technology categories.
The following observations can be made from the two level Geo-Spatial patent citation network portrayed
in figure 3.
1. The tech areas contributing to GaN based LED invention are (From the backward citations)-
a) Semiconductors
b) Photolithography

c) Optics
d) Domestic Appliances
2. The prominent sub-technologies that spun out of this invention/technology are (From forwards)-
a) Semiconductors
b) Drugs & Med Chemicals
c) Photolithography
d) Optics
e) Lighting
3. Other sub-technologies that make use of the Blue LED invention, that show a significant presence
in the graph (From forwards) are-
a) Heating & Cooling
b) Chemicals & Polymers
c) Plastics & Wheels
d) Vehicles
e) Cosmetics & Med Chemicals
f) Lab Equipment
g) Metals
h) Measurement
i) Computing
4. The above graph also helps in understanding the point of time a particular technology or
application contributed to or spun out from the invention being considered. For instance, we can
observe that the use of the Blue LED technology in Drugs and Med Chemicals started since 1998.
This seminal patent led to the filing of relatively more important patents in the later years (The bigger
nodes to the right of the seminal patent). In the above layout, the importance of a patent is defined based
on the number of forward citations a patent has accumulated. The size of a node is proportional to the
number of forward citations it has accumulated.
In figure 4, the edges are weighted based on the similarity of the first IPC class code of the patents
connected to each other. Other means of weighting the edges could be based on the number of common
patent classes between the two patents or based on the number of common citations between them.

Figure 4. The patent citation network of the seminal patent in a Geo-Spatial mode using IPC based NBER
categories with weighted edges connecting similar patents
Figure 4 illustrates the two level patent citations rendered using the Geo-Spatial layout along with
weighted edges indicating patent similarity. The edge colors are same as the color of the origin node
indicating source of knowledge flow.
The following observations can be made from the graph illustrated in figure 4.
1. The patents that are relatively similar are linked with thick edges.
2. Apart from many citations emanating from the patents pertaining to the Semiconductor sub-
technology, a considerable amount of references are also from the patents pertaining to the
‘Drugs and Med Chemicals’ sub-technology and thus indicating the importance of this sub-
technology in this analysis.

Figure 5. The patent citation network of the seminal patent in a Geo-Spatial mode using Relecura
Technology categories
Figure 5 illustrates the two level patent citation network in the Geo-Spatial mode with the size of the node
indicating the Relecura Patent Quality of the patent and the color codes representing various Relecura
Technology categories.
Table 2. Various Relecura Technology categories in the patent citation network
Vehicle lighting & signalling Other Lighting Devices Electric Heating and Lighting
Therapeutic devices - Energy Optical property modification
devices
Electric Elements -
Semiconductor devices
Testing materials Optical elements Electric elements - LASERS
Sterilization, Dressing Non portable lighting devices Displays
Sealing materials and drilling
fluids
Nano structure applications Dentistry - Oral Care
Printed Circuits Metal Coating Crystal Growth
Plastics Shaping Measurement - Temperature Chemical/Physical processes -
catalysis

Pictorial Communication Measurement - Light Carbocyclic compounds
Photography - Camera Light Sources - Others Design patents
Figure 5 exemplifies the milieu of the seminal patent in terms of Relecura Technology categories.
Figure 6 illustrates the two level patent citations rendered using the Geo-Spatial layout along with edge
weights indicating patent similarity. The edge colors are same as the color of the origin node indicating
source of knowledge flow. The size of the node indicates the quality of the patent in terms of Relecura
Patent Quality and the colors represent various Relecura Technology categories.
Figure 6. The patent citation network of the seminal patent in a Geo-Spatial mode using Relecura
Technology categories with weighted edges connecting similar patents
Figure 6 is dominated with edges emanating from the patents pertaining to the category - Electric
elements – Semiconductor devices. It is worth noting that a significant number of links also emerged from
areas viz. Sealing materials and drilling fluids; Nanostructure applications; Dentistry – Oral care and
Crystal growth.
Conclusions:
A holistic picture of a technology domain can be obtained by analyzing the citation network of a seminal
patent. This methodology can be primarily used to determine what sub-technologies contributed to the
technology in question, at what period in time and with how much impact. The same can also be extended
to identify what sub technologies spun out from the technology/invention and also in identifying various

application areas of the technology in question. At a higher level, it can be used to understand the
development and flow of scientific knowledge in a particular field.
References:
1. A.B. Jaffe, Evidence from patents and patent citations on the impact of NASA and other federal
labs on commercial innovation, J. Ind. Econ. 46 (2) (1998) 183–205.
2. A.B. Jaffe, M. Trajtenberg, International knowledge flows: evidence from patent citations, Econ.
Innov. New Technol. 8 (1999) 105–136.
3. A.B. Jaffe, M. Trajtenberg, R. Henderson, Geographic localization of knowledge spillovers as
evidenced by patent citations, Q. J. Econ. 108 (1993) 578–598.
4. Bastian M., Heymann S., Jacomy M. (2009). Gephi: an open source software for exploring and
manipulating networks. International AAAI Conference on Weblogs and Social Media.
5. Börner, K. (2012). Analyzing and Visualizing Science.
6. Buchanan, B., & Corken, R. (2010). A toolkit for the systematic analysis of patent data to assess a
potentially disruptive technology, (c).
7. Chen, S., Huang, M., & Chen, D. (2012). Technological Forecasting & Social Change Identifying
and visualizing technology evolution: A case study of smart grid technology. Technological
Forecasting & Social Change, 79(6), 1099–1110.
8. Érdi, Péter, Kinga Makovi, Zoltán Somogyvári, Katherine Strandburg, Jan Tobochnik, Péter Volf,
and László Zalányi. 2013. “Prediction of emerging technologies based on analysis of the US
patent citation network.” Scientometrics 95, no. 1 (April): 225-42.
9. Geum, Y., Lee, S., Yoon, B., & Park, Y. (2013). Technovation Identifying and evaluating
strategic partners for collaborative R & D: Index-based approach using patents and publications.
Technovation, 33(6-7), 211–224.
10. Gress, B. (2010). Properties of the USPTO patent citation network: 1963–2002. World Patent
Information, 32(1), 3–21.
11. Hall, B. H., A. B. Jaffe, and M. Trajtenberg (2001). "The NBER Patent Citation Data File:
Lessons, Insights and Methodological Tools." NBER Working Paper 8498.
12. Holloway, P. & McGuire, G. (1995). Handbook of compound semiconductors growth,
processing, characterization, and devices. Park Ridge, N.J., U.S.A: Noyes Publications.
13. National Bureau of Economic Research (2014), http://www.nber.org/patents/ (accessed on June,
2014)
14. Relecura (2014), http://www.relecura.com (accessed on June, 2014)
15. Aistemos(2014),https://aistemos.com/wp-content/uploads/2014/07/CIPHER-Briefing_Patent-
Strength_July20141.pdf (accessed on October, 2014)
16. Jacomy M, Venturini T, Heymann S, Bastian M (2014) ForceAtlas2, a Continuous Graph Layout
Algorithm for Handy Network Visualization Designed for the Gephi Software. PLoS ONE 9(6):
e98679. doi:10.1371/journal.pone.0098679
17. Kim C., Lee, H., Seol, H., & Lee, C. (2011). Expert Systems with Applications Identifying core
technologies based on technological cross-impacts: An association rule mining (ARM) and
analytic network process (ANP) approach. Expert Systems with Applications, 38(10), 12559–
12564.
18. Kobourov S G., “Force-directed drawing algorithms”, Handbook of Graph Drawing and
Visualization, Tamassia R. ed., CRC Press, August 2013.
19. Lee C., Jeon J., & Park Y. (2011). Technological Forecasting & Social Change Monitoring trends
of technological changes based on the dynamic patent lattice: A modified formal concept analysis
approach. Technological Forecasting & Social Change, 78(4), 690–702.
20. Lee S., Yoon B., Lee C., & Park J. (2009). Business planning based on technological capabilities:
Patent analysis for technology-driven road mapping. Technological Forecasting and Social
Change, 76(6), 769–786.

21. Marks, P. Patent trawler aims to predict next hot technologies. New Scientist. Retrieved from
http://is.gd/EMIEkU
22. Maruska, P. (1990). A BRIEF HISTORY OF GaN BLUE LIGHT-EMITTING DIODES.
23. Meister, C., & Meister, M. (2005). Trends and trajectories in mems-related technologies: an
analysis on the basis of patent application data. CAS 2005 Proceedings. 2005 International
Semiconductor Conference, 2005, 1, 187–190.
24. Nakamura, S., & Krames, M. R. (2013). History of Gallium–Nitride-Based Light-Emitting
Diodes for Illumination. Proceedings of the IEEE, 101(10), 2211–2220.
25. Park, H., Kim, K., Choi, S., & Yoon, J. (2013). Expert Systems with Applications A patent
intelligence system for strategic technology planning. Expert Systems with Applications, 40(7),
2373–2390.
26. Park, H., Yoon, J., & Kim, K. (2013). Expert Systems with Applications Using function-based
patent analysis to identify potential application areas of technology for technology transfer.
Expert Systems with Applications, 40(13), 5260–5265.
27. Patel, D., & Ward, M. R. (2011) Using patent citation patterns to infer innovation market
competition. Research Policy, 40, 886-894.
28. Porter, A. L. (2005). QTIP: Quick technology intelligence processes. Technological Forecasting
and Social Change, 72(9), 1070–1081.
29. Porter, A., Lucian. K. (2013). Patent Overlay Mapping - Visualizing Tech Distance. JASIST,
(August).
30. Thorleuchter, D., Poel, D. Van Den, & Prinzie, A. (2010). A compared R & D-based and patent-
based cross impact analysis for identifying relationships between technologies. Technological
Forecasting & Social Change, 77(7), 1037–1050.
31. Xiang, X., Cai, H., Lam, S., & Pei, Y. (2013). International knowledge spillover through co-
inventors: An empirical study using Chinese assignees' patent data. Technological Forecasting
and Social Change, 161-174.
32. Yoon, B., & Park, Y. (2004). A text-mining-based patent network: Analytical tool for high-
technology trend, 15, 37–50.

Appendix:
Biologics Domestic appliances Machine Tools Radio, Comm
Catalysis & Separation Drugs, Med Chem Measurement Recording
Chem& Polymers Electric Power Med Instruments Semiconductors
Combustion Engines Food Medical devices Telephone Comm
Computing Furnace Metals Textiles
Construction Heating & Cooling Null Turbines & Engines
Copying & Printing Info Transmission Optics TV, Imaging &Comm
Cosm& Med Chem Lab equipment Photolithography Vehicle parts
Data Commerce Lighting Plastics & Wheels Vehicles
NBER IPC based Technologies
Vehicle lighting & signalling Other Lighting Devices Electric Heating and Lighting
Therapeutic devices - Energy Optical property modification devices Electric Elements - Semiconductor devices
Testing materials Optical elements Electric elements - LASERS
Sterilization, Dressing Non portable lighting devices Displays
Sealing materials and drilling fluids Nano structure applications Dentistry - Oral Care
Printed Circuits Metal Coating Crystal Growth
Plastics Shaping Measurement - Temperature Chemical/Physical processes - catalysis
Pictorial Communication Measurement - Light Carbocyclic compounds
Photography - Camera Light Sources - Others Design patents
Relecura Technology Categories used in study

Technology analysis using patent citation network of a seminal patent final

More Related Content

What's hot (8)

Similar to Technology analysis using patent citation network of a seminal patent final (20)

More from Shanmukha S. Potti (8)

Recently uploaded (20)

Technology analysis using patent citation network of a seminal patent final