Scientometrics (2014) 99:2735

DOI 10.1007/s11192-013-1079-4

A routine for measuring synergy in universityindustry

government relations: mutual information as a Triple-
Helix and Quadruple-Helix indicator

Loet Leydesdorff Han Woo Park Balazs Lengyel

Received: 28 June 2013 / Published online: 10 July 2013

 Akademiai Kiado, Budapest, Hungary 2013

Abstract Mutual information in three (or more) dimensions can be considered as a

Triple-Helix indicator of possible synergy in universityindustrygovernment relations. An
open-source routine th4.exe makes the computation of this indicator interactively available
at the internet, and thus applicable to large sets of data. Th4.exe computes all probabilistic
entropies and mutual information in two, three, and, if available in the data, four dimen-
sions among, for example, classes such as geographical addresses (cities, regions), tech-
nological codes (e.g. OECDs NACE codes), and size categories; or, alternatively, among
institutional addresses (academic, industrial, public sector) in document sets. The relations
between the Triple-Helix indicatoras an indicator of synergyand the Triple-Helix
model that specifies the possibility of feedback by an overlay of communications, are also
discussed.

Keywords Indicator  Triple Helix  Quadruple Helix  Software  Information theory 

Mutual information  Mutual redundancy

L. Leydesdorff (&)
Amsterdam School of Communication Research (ASCoR), University of Amsterdam,
Kloveniersburgwal 48, 1012 CX Amsterdam, The Netherlands
e-mail: [email protected]
URL: http://www.leydesdorff.net

H. W. Park
Department of Media and Communication, YeungNam University, 214-1, Dae-dong, Gyeongsan-si,
Gyeongsangbuk-do 712-749, South Korea
e-mail: [email protected]

B. Lengyel
International Business School Budapest, Tarogato ut 2-4, Budapest 1021, Hungary
e-mail: [email protected]

123
28 Scientometrics (2014) 99:2735

Introduction

The Triple Helix model of universityindustrygovernment relations (Etzkowitz and

Leydesdorff 1995, 2000) rapidly became interesting to policy makers because one can also
interpret this program of studies as a call for greater collaboration locally. However, both
the sciences and markets develop globally. The task of government is to retain wealth from
knowledge or knowledge from wealth (e.g. oil revenues) by creating institutional frame-
works which enable agents (firms, scholars) in regions and nations to be both global and
local players, but with potentially different roles (Freeman and Perez 1988). Thus, a trade-
off between international and national/regional collaboration and orientation is required
which is more difficult than a simple recipe in either direction. Some national systems, for
example, are insufficiently open to global developments, while others are perhaps not
sufficiently integrated internally. Note that the use of an indicator always requires theo-
retical interpretation.
Case studies inform us about best practices and one can learn from failures, but the
results remain at the phenomenological level; the phenotypes of the Triple Helix
partners (university, industry, government) can be expected to carry genotypical functions
such as novelty production (in science), wealth generation (on the market), and governance
at different scales (Ivanova and Leydesdorff 2012; cf. Hodgson and Knudsen 2011). From
the perspective of the Triple-Helix theory, an indicator should ideally enable us to dis-
tinguish between a vicious lock-in into existing densities of relations (Biggiero 1998) and
too looseor footloose (Vernon 1979)relationships at the global level that are
insufficiently embedded locally (Cohen and Levinthal 1990). Furthermore, such an indi-
cator should not predicate specific levels of integration (at the national or regional level),
but make this question empirical. The indicator should allow for extension to more than
three dimensions. In other words, systemness can be considered as a hypothesis that can
be tested in terms of the data. Is more systemness developed locally, regionally, sec-
torially, or nationally; and in which dimensions? (Carlsson 2006).
We wish to contribute to this special issue about Mapping Triple Helix Innovation for
Developing and Transitional Economies: Webometrics, Scientometrics, and Informetrics,
by making a routine available as open-source at the Internet for the computation of mutual
information in three (or four) dimensions. Mutual information in three (or more) dimen-
sions can be positive or negative and thus indicate synergy or a lack thereof in Triple-Helix
(or Quadruple Helix) data. Increased access to large data sets, in our opinion, makes it
urgent to develop tools that can be used flexibly to analyze data in different dimensions and
using different units of analysis (such as firm data, publication data, etc.; cf. Park 2010;
Leydesdorff et al. 2012). The Triple-Helix indicator emerged from discussions about
Triple-Helix relations among universities, industries, and government agencies, but its use
is in no way confined to such relations alone. In general, one can use this indicator for
testing the emergence of next-order systemness in the case of autocatalysis among
carrying agents at the next-lower level (e.g. Ulanowicz 2009). The Triple-Helix overlay
(Etzkowitz and Leydesdorff 2000) can be considered as such a potentially emerging next-
level system.

The Triple Helix indicator of synergy in universityindustrygovernment relations

Building on Ulanowicz (1986), at p. 143) proposal to use mutual information in three

dimensions as an indicator of next-order systemness among three (or more) independent
dimensions in the data, Leydesdorff (2003) suggested using university, industrial, and

123
Scientometrics (2014) 99:2735 29

public-sector addresses as provided by the Science Citation Index, and thus to distinguish
among nations in terms of the synergy in these institutional relations. Synergy can then be
considered as a reduction of uncertainty in universityindustrygovernment relations.
Two of the three dimensions can be partially and/or spuriously correlated in the third
dimension. Spurious correlation can reduce uncertainty without being visible in the data
without analysisand therefore latent; whereas partial correlation can be measured using
statistics such as Pearsons rxy|z, or other measures such as Shannons Hxy|z as a measure of
uncertainty in two dimensions conditioned on a third, and the corresponding mutual
information Txy|z.
Mutual information in more than two dimensions can be derived from Shannons (1948)
formulas (e.g. Yeung 2008, at pp. 59f.), but it can no longer be considered as a Shannon
entropy (Krippendorff 2009a, b; Leydesdorff 2010). Its value can be positive, negative or
zero, whereas Shannon types of information are necessarily positive (Theil 1972, pp. 59f.).
It can be shown (e.g. Abramson 1963, at pp. 128 ff.; McGill 1954) that the mutual
information in or transmission (T) among three dimensions (Txyz) is equal to:
TXYZ HX HY HZ  HXY  HXZ  HYZ HXYZ 1
P P
in which formula HX  pX logpX and HXY  pXY logpXY , etc., for one or
more discrete variables X and Y, etc. Using the two-base for the logarithm, all these
uncertainty measures are expressed in bits of information.
The further extension to more than three dimensions is straightforward. For example,
Leydesdorff and Sun (2009); cf. Kwon et al. 2012) considered the mutual information in
the four dimensions of university (u), industry (i), government (g), or foreign (f) co-
authorship relations in the case of Japan, using:
Tuigf Hu Hi Hg Hf  Hui  Hug  Huf  Hig  Hif  Hgf
2
Huig Huif Hugf Higf  Huigf
These scientometric studies (see also: Park et al. 2005; Park and Leydesdorff 2010; Ye
et al. 2012) considered publications as units of analysis. A series of econometric studies
was also developed using firms as units of analysis and firm size, technological classifi-
cation, and address information as proxies for the economic, technological, and govern-
mental dimensions, respectively. These studies comprise: Leydesdorff et al. (2006) about
The Netherlands; Leydesdorff and Fritsch (2006) about Germany; Lengyel and Ley-
desdorff (2011) about Hungary; Strand and Leydesdorff (2013) about Norway; Ley-
desdorff and Strand (2012) about Sweden; and Perevodchikov et al. (2013) about the
Russian Federation.

Automating the Triple-Helix indicator

The first version of an automated routine that allows users online to compute values for the
TH Indicatorthat is, mutual information in three dimensionswas made in 2008 and is
available at http://www.leydesdorff.net/th.
In 2009, this routine was elaborated with Krippendorffs (1980) IABC?AB,AC,BC as a
measure for the uncertainty that is added by a three-dimensional interaction to the sum
total of information contents of the two-dimensional interactions. Krippendorffs IABC?-
AB,AC,BC, however, is a Shannon-type information that is necessarily positive (Leydesdorff
2010, 2011). One can also use Occam 3 for its computation (at http://occam.research.
pdx.edu/occam/weboccam.cgi?action=search).

123
30 Scientometrics (2014) 99:2735

Following Krippendorff (2009b), the TH indicatorthat is, mutual information in three

dimensionscan be considered as the difference between the redundancy generated in the
loops of communication possible in three-way interactions, and the (Shannon-type)
information generated in these interactions. If the TH indicator is negative in terms of (e.g.
bits of) information, this indicates a surplus of redundancy being generated. From the
perspective of the TH model, this redundancythat is, a relative reduction of uncer-
taintycan also be considered as a measure of synergy at the systems level. In the TH
model, this is considered as an overlay of communications which can feed back as a
hyper-cycle on top of the cycles of communication within each of the helices. The hyper-
cycle is a consequence of interactions among the relations between the underlying systems.
In other words, if universities and industries are already collaborating, the role of
government can be different from a configuration where the two spheres operate at a
distance. In a system with three functions, the order of operations can also be expected to
make a difference: if government takes the lead, the result is different from leaving the lead
to industry (Ivanova and Leydesdorff 2012).
The first routine (TH.exe) can be used for single values of the seven relevant variables:
U, I, G, UI, UG, IG, and UIG. In most cases, these will be count data in documents (or
other) sets. The new routine th4.exe (at http://www.leydesdorff.net/software/th4) provides
the user with the option to use batches of cases and up to four independent dimensions
(Carayannis and Campbell 2009, 2010; Leydesdorff 2012a, b). Each case (for example, a
publication or firm) is represented as a single line. Each line first describes the case using
an identifier (e.g., a sequence number), a value for potentially four variables called w, x, y,
and z. If the fourth variable z is not specified, a single TH between w, x, and y is computed
and all combinations with z are set equal to zero.
The variables and identifiers are considered as nominal variables. An input file has to be
called data.txt and can, for example, be shaped as follows:
In the case of universityindustrygovernment relations, for example, one may wish to
code the second variable Y when a university address is present and N when not; and,
mutatis mutandis, the third variable for industry addresses, etc. This file (data.txt) should
be stored in the same folder as th4.exe. The latter program generates a file th4.dbf in which
the values for all relevant parameters are stored (Table 2). (The dbf-file can be read into
such programs as Excel, SPSS, or OpenOffice.)
For example, the second variable (that is, the first dimension w) in Table 1 has three
times the value of 1, and once a 2. The uncertainty H(w) in this distribution is
consequentially -(3/4) log2 (3/4) - (1/4) log2 (1/4) = 0.31 ? 0.50 = 0.81 bits of infor-
mation. Table 2 shows furthermore that two of the three possible mutual informations in
three dimensions are negative in this case, and the four-dimensional T(wxyz) would also be
-0.19 bits.
This above example is fictitious. Table 3 provides the data of six among 14,552 firms in
the Trondheim (Tromso) region with their firm identity number, geographical code,
NACE-code, and size code, as used by Strand and Leydesdorff (2013).

Table 1 Fictitious example of

id1, 1, b, region1, 2
input data for th4.exe
id2, 2, a, region2, 1
id3, 1, a, region2, 2
id4, 1, b, region5, 1

123
Scientometrics (2014) 99:2735 31

Table 2 Output of th4.exe on the

Entropy and transmission values in bits of information
basis of the data in Table 1, in bits
of information
H(W) 0.81
H(X) 1.00
H(Y) 1.50
H(Z) 1.00
H(WX) 1.50
H(WY) 2.00
H(WZ) 1.50
H(XY) 1.50
H(XZ) 2.00
H(YZ) 2.00
H(WXY) 2.00
H(WXZ) 2.00
H(WYZ) 2.00
H(XYZ) 2.00
H(WXYZ) 2.00
T(WX) 0.31
T(WY) 0.31
T(WZ) 0.31
T(XY) 1.00
T(XZ) 0.00
T(YZ 0.50
T(WXY) 0.31
T(WXZ) -0.19
T(WYZ) -0.19
T(XYZ) 0.00
T(WXYZ) -0.19

Table 3 Six firms in the Tromso

459695,1901,5,3
region (coded 1901) with their
identity-codes in the first column, 459696,1901,5,5
and NACE-codes (OECD) and 459697,1901,11,1
size codes in columns 3 and 4,
459698,1901,11,2
respectively
459699,1901,11,2
459700,1901,11,2

When one feeds the entire set of 14,552 firms into th4.exe, the resulting value of
T(wxy) = -0.135 bits of information. Note that the fourth variable z was not specified in
this set. As Table 3 indicates, the quotation marks are not strictly needed, but they are
convenient in order to remember the nominal character of this data. In the future, we plan
to make a similar program for numerical data, but in many applications the data at the case
level will be coded into categories.

123
32 Scientometrics (2014) 99:2735

The output is collected in a file named th4.dbf that can be read into Excel, SPSS, or
OpenOffice. If this file does not yet exist, it is generated by the routine; otherwise, a new
record is added each time this routine is used. Thus, one can run subsets (e.g. regions)
consecutively (by renaming files into data.txt in the right order), and each region will be
represented as a row in the file th4.dbf.

Conclusions and discussion

The Triple-Helix indicator fulfills the requirements that were specified in the introduction
on the basis of the Triple-Helix model: the trade-off between a relative rigidity which may
emerge in an institutional network that is too densely populated with links, versus too
much volatility in a network that is internationally oriented, can be quantified in terms of
this signed information measure. Furthermore, there are virtually no size limitations for the
data. (The current program is limited to 2 GByte of data so that it is compatible with both a
32-bits and 64-bits operating system). However, the interpretation remains a bit elusive: a
negative information is a redundancy. This redundancy is generated in next-order loops in
the information processing which remain hypothesized, but cannot be observed directly
(Luhmann 1984, 1995, p. 226; p. 164). Figure 1
In other words, one can hypothesize the knowledge base of an economy to be structured
at the national, regional, or other levels, and then test these hypotheses in terms of the
Triple-Helix indicator. Figure 2, for example, provides such a map for Sweden. The results
can be mapped onto the geography after decomposition (contributions to the synergy are
indicated as -DTTGO in Fig. 2), but one can also decompose in the other dimensions (Theil

Fig. 1 The routine TH.exe available at http://www.leydesdorff.net/th

123
Scientometrics (2014) 99:2735 33

Fig. 2 Contributions to the reduction of uncertainty at the level of 21 Swedish counties. (DT among
technology, geography, and organization; Source: Leydesdorff and Strand 2012)

1972), for example, between high-tech manufacturing and knowledge-intensive services

(e.g. Leydesdorff et al. 2006) or between national and international collaborations (Ley-
desdorff and Sun 2009). The test-statistics in terms of significance are as yet poorly
developed for entropy statistics. Although one is able to relate information theory to
likelihood ratios between alternative hypotheses and Bayesian statistics (Sheskin 2011, at
pp. 384 ff.; Leydesdorff 1992, 1995), the major application of information theory remains
oriented towards descriptive statistics of evolving systems (Leydesdorff 2012a, b).
The further extension to more than three dimensions is popular among scholars who
work with the Triple-Helix model (e.g. Bunders et al., 1999; Carayannis and Campbell
2009, 2010; Leydesdorff and Etzkowitz 2003). Equations (1) and (2) show that the sign
may change, when one extends from three to four (or more) dimensions. Krippendorff
(2009b), at p. 670) formalized this in terms of the cardinality of a parameter C in a more
difficult summary equation. However, the interpretation of these next-order terms is dif-
ferent (Leydesdorff 2012a, b; Leydesdorff and Ivanova 2013).
Nevertheless, we hope to have provided a service to the community of Triple-Helix
researchers by making the source code available on the internet and as open source (at
http://www.leydesdorff.net/software/th4/th4.prg). One can also use this routine to check
ones own calculations, for example, in Excel. Different from comprehensive programs for
science mapping, these routines are meant for exploration and research (Cobo et al. 2011).
Ivan Cucco (personal communication, April 14, 2013), for example, recompiled the pro-
gram under Linux using the (freeware) Harbour platform. Other feedback from the side of
users is appreciated and will hopefully lead to further improvements.

Acknowledgments We thank ivind Strand for providing data about Tromso. We acknowledge support
from the SSK (Social Science Korea) Program funded by the National Research Foundation of South Korea;
NRF-2010-330-B00232; Balazs Lengyel acknowledges support from the Hungarian Scientific Research
Fund (PD106290).

123
34 Scientometrics (2014) 99:2735

References

Abramson, N. (1963). Information theory and coding. New York: McGraw-Hill.

Biggiero, L. (1998). Italian industrial districts: A Triple Helix pattern of problem solving. Industry and
Higher Eductation, 12(4), 227234.
Bunders, J. F. G., Broerse, J. E. W., & Zweekhorst, M. B. M. (1999). The Triple Helix enriched with the user
perspective: A view from Bangladesh. The Journal of Technology Transfer, 24(2), 235246.
Carayannis, E. G., & Campbell, D. F. J. (2009). Mode 3 and Quadruple Helix: toward a 21st century
fractal innovation ecosystem. International Journal of Technology Management, 46(3), 201234.
Carayannis, E. G., & Campbell, D. F. J. (2010). Triple Helix, Quadruple Helix and Quintuple Helix and how
do knowledge, innovation, and environment relate to each other? International Journal of Social
Ecology and Sustainable Development, 1(1), 4169.
Carlsson, B. (2006). Internationalization of innovation systems: A survey of the literature. Research Policy,
35(1), 5667.
Cobo, M. J., Lopez-Herrera, A. G., Herrera-Viedma, E., & Herrera, F. (2011). Science mapping software
tools: Review, analysis, and cooperative study among tools. Journal of the American Society for
Information Science and Technology, 62(7), 13821402.
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and inno-
vation. Administrative Science Quarterly, 35(1), 128153.
Etzkowitz, H., & Leydesdorff, L. (1995). The Triple Helixuniversityindustrygovernment relations: A
laboratory for knowledge-based economic development. EASST Review, 14, 1419.
Etzkowitz, H., & Leydesdorff, L. (2000). The dynamics of innovation: From national systems and Mode 2
to a Triple Helix of universityindustrygovernment relations. Research Policy, 29(2), 109123.
Freeman, C., & Perez, C. (1988). Structural crises of adjustment, business cycles and investment behaviour.
In G. Dosi, C. Freeman, R. Nelson, G. Silverberg, & L. Soete (Eds.), Technical change and economic
theory (pp. 3866). London: Pinter.
Hodgson, G., & Knudsen, T. (2011). Darwins conjecture: The search for general principles of social and
economic evolution. Chicago/London: University of Chicago Press.
Ivanova, I. A., & Leydesdorff, L. (2012). Rotational symmetry and the transformation of innovation systems
in a Triple Helix of universityindustrygovernment relations; available at http://arxiv.org/abs/
1211.2573.
Krippendorff, K. (1980). Q; an interpretation of the information theoretical Q-measures. In R. Trappl, G.
J. Klir, & F. Pichler (Eds.), Progress in cybernetics and systems research (Vol. VIII, pp. 6367). New
York: Hemisphere.
Krippendorff, K. (2009a). W. Ross Ashbys information theory: A bit of history, some solutions to prob-
lems, and what we face today. International Journal of General Systems, 38(2), 189212.
Krippendorff, K. (2009b). Information of interactions in complex systems. International Journal of General
Systems, 38(6), 669680.
Kwon, K. S., Park, H. W., So, M., & Leydesdorff, L. (2012). Has globalization strengthened South Koreas
national research system? National and international dynamics of the Triple Helix of scientific co-
authorship relationships in South Korea. Scientometrics, 90(1), 163175. doi:10.1007/s11192-011-
0512-9.
Lengyel, B., & Leydesdorff, L. (2011). Regional innovation systems in Hungary: The failing synergy at the
national level. Regional Studies, 45(5), 677693. doi:10.1080/00343401003614274.
Leydesdorff, L. (1992). Knowledge representations, bayesian inferences, and empirical science studies.
Social Science Information, 31(2), 213237.
Leydesdorff, L. (1995). The challenge of scientometrics: The development, measurement, and self-orga-
nization of scientific communications. Leiden: DSWO Press, Leiden University.
Leydesdorff, L. (2003). The mutual information of universityindustrygovernment relations: An indicator
of the Triple Helix dynamics. Scientometrics, 58(2), 445467.
Leydesdorff, L. (2010). Redundancy in systems which entertain a model of themselves: Interaction infor-
mation and the self-organization of anticipation. Entropy, 12(1), 6379. doi:10.3390/e12010063.
Leydesdorff, L. (2011). Structuration by intellectual organization: The configuration of knowledge in
relations among scientific texts. Scientometrics, 88(2), 499520.
Leydesdorff, L. (2012a). The Triple Helix, Quadruple Helix,, and an N-tuple of helices: Explanatory
models for analyzing the knowledge-based economy? Journal of the Knowledge Economy, 3(1),
2535. doi:10.1007/s13132-13011-10049-13134.
Leydesdorff, L. (2012b). Statistics for the dynamic analysis of scientometric data: The evolution of the
sciences in terms of trajectories and regimes. Scientometrics,. doi:10.1007/s11192-012-0917-0.

123
Scientometrics (2014) 99:2735 35

Leydesdorff, L., Dolfsma, W., & Van der Panne, G. (2006). Measuring the knowledge base of an economy
in terms of Triple-Helix relations among Technology, Organization, and Territory. Research Policy,
35(2), 181199.
Leydesdorff, L., & Etzkowitz, H. (2003). Can The Public be considered as a fourth Helix in university
industrygovernment relations? Report of the fourth Triple Helix conference. Science and Public
Policy, 30(1), 5561.
Leydesdorff, L., & Fritsch, M. (2006). Measuring the knowledge base of regional innovation systems in
Germany in terms of a Triple Helix dynamics. Research Policy, 35(10), 15381553.
Leydesdorff, L., & Ivanova, I. A. (2013). Mutual redundancies in inter-human communication systems:
Steps towards a calculus of processing meaning. Journal of the American Society for Information
Science and Technology; available at http://arxiv.org/abs/1301.6849.
Leydesdorff, L., Rotolo, D., & de Nooy, W. (2012). Innovation as a nonlinear process, the scientometric
perspective, and the specification of an Innovation Opportunities Explorer. Technology Analysis and
Strategic Management; available at http://arxiv.org/abs/1202.6235.
Leydesdorff, L., & Strand, . (2012). The Swedish system of innovation: Regional synergies in a knowl-
edge-based economy. Journal of the American Society for Information Science and Technology,
62(11), 21332146.
Leydesdorff, L., & Sun, Y. (2009). National and international dimensions of the Triple Helix in Japan:
Universityindustrygovernment versus international co-authorship relations. Journal of the American
Society for Information Science and Technology, 60(4), 778788.
Luhmann, N. (1984). Soziale systeme. Grundri einer allgemeinen theorie. Frankfurt: Suhrkamp.
Luhmann, N. (1995). Social systems. Stanford, CA: Stanford University Press.
McGill, W. J. (1954). Multivariate information transmission. Psychometrika, 19(2), 97116.
Park, H. W. (2010). Mapping the e-science landscape in South Korea using the webometrics method.
Journal of Computer-Mediated Communication, 15(2), 211229.
Park, H. W., Hong, H. D., & Leydesdorff, L. (2005). A comparison of the knowledge-based innovation
systems in the economies of South Korea and The Netherlands using Triple Helix indicators.
Scientometrics, 65(1), 327.
Park, H. W., & Leydesdorff, L. (2010). Longitudinal trends in networks of universityindustrygovernment
relations in South Korea: The role of programmatic incentives. Research Policy, 39(5), 640649.
Perevodchikov, E., Uvarov, A., & Leydesdorff, L. (2013). Measuring synergy in the Russian innovation
system. Paper presented at the 12th international conference about the triple helix of university-
industry-government relations, London, UK.
Shannon, C. E. (1948). A mathematical theory of communication. Bell System Technical Journal, 27,
379423 and 623656.
Sheskin, D. J. (2011). Handbook of parametric and nonparametric statistical procedures (5th ed.). Boca
Raton, FL: Chapman and Hall/CRC.
Strand, ., & Leydesdorff, L. (2013). Where is synergy in the Norwegian innovation system indicated?
Triple Helix relations among technology, organization, and geography. Technological Forecasting and
Social Change, 80(3), 471484.
Theil, H. (1972). Statistical decomposition analysis. Amsterdam/London: North-Holland.
Ulanowicz, R. E. (1986). Growth and development: Ecosystems phenomenology. San Jose: toExcel.
Ulanowicz, R. E. (2009). A third window: Natural life beyond Newton and Darwin. West Conshohocken:
Templeton Foundation Press.
Vernon, R. (1979). The product cycle hypothesis in a new international environment. Oxford Bulletin of
Economics and Statistics, 41(4), 255267.
Ye, F. Y., Yu, S. S., & Leydesdorff, L. (2012). The Triple Helix of universityindustrygovernment
relations at the country level, and its dynamic evolution under the pressures of globalization; available
at http://arxiv.org/abs/1209.1260.
Yeung, R. W. (2008). Information theory and network coding. New York, NY: Springer.

123