Professional Documents
Culture Documents
Download textbook Complex Networks Amp Their Applications V Proceedings Of The 5Th International Workshop On Complex Networks And Their Applications Complex Networks 2016 1St Edition Hocine Cherifi ebook all chapter pdf
Download textbook Complex Networks Amp Their Applications V Proceedings Of The 5Th International Workshop On Complex Networks And Their Applications Complex Networks 2016 1St Edition Hocine Cherifi ebook all chapter pdf
https://textbookfull.com/product/complex-networks-their-
applications-xii-volume-2-1st-edition-hocine-cherifi/
https://textbookfull.com/product/complex-networks-principles-
methods-and-applications-1st-edition-vito-latora/
Malware Diffusion Models for Wireless Complex Networks.
Theory and Applications 1st Edition Karyotis
https://textbookfull.com/product/malware-diffusion-models-for-
wireless-complex-networks-theory-and-applications-1st-edition-
karyotis/
https://textbookfull.com/product/reconfigurable-cellular-neural-
networks-and-their-applications-mustak-e-yalcin/
https://textbookfull.com/product/complex-networks-in-software-
knowledge-and-social-systems-milos-savic/
https://textbookfull.com/product/analysis-and-control-of-output-
synchronization-for-complex-dynamical-networks-jin-liang-wang/
https://textbookfull.com/product/complex-systems-design-amp-
management-proceedings-of-the-tenth-international-conference-on-
complex-systems-design-management-csd-m-paris-2019-guy-andre-boy/
Studies in Computational Intelligence 693
Hocine Cherifi
Sabrina Gaito
Walter Quattrociocchi
Alessandra Sala Editors
Complex
Networks & Their
Applications V
Proceedings of the 5th International
Workshop on Complex Networks and
their Applications (COMPLEX NETWORKS
2016)
Studies in Computational Intelligence
Volume 693
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
About this Series
Editors
123
Editors
Hocine Cherifi Walter Quattrociocchi
University of Burgundy IMT Lucca
Dijon Lucca
France Italy
The International Workshop on Complex Networks & their Applications was first
held in 2012. It was initially conceived as a forum to bring together researchers from
a wide variety of fields ranging from Computational Social Science, to Economic
Complexity, up to Bioinformatics to review current scientific work and formulate new
directions in network science. The tradition has continued with an annual single-track
meeting that has become one of the leading international events in the field. Fuelled
by the skills and expertise of participants from these diverse research fields, this
workshop allows for cross-fertilization between fundamental and applied research. It
offers a unique opportunity for reflection on the current state of the field, unanswered
but critical questions, and potential future directions.
This volume of proceedings provides an opportunity for readers to engage with
a selection of papers presented during the Fifth edition, hosted by the University
of Milan (Italy), from November 30 to December 02, 2016. Although, they do not
provide a fully comprehensive coverage of the field, the 65 papers selected by the
Scientific Committee reflect the interdisciplinary nature of the scientific areas covered
by the workshop. They have been organized in 11 sections reflecting multiple aspects
of complex network research:
• Network models
• Network measures
• Community structure
• Network dynamics
• Diffusion, epidemics and spreading processes
• Resilience and control
• Network visualization
• Social and political networks
• Networks in finance and economics
• Biological and ecological networks
• Network analysis
A very encouraging response has been received by COMPLEX NETWORKS 2016
in terms of submissions. The 204 contributions that we received from 47 countries
v
vi Preface
around the world reflect the great vitality and diversity of the complex network
community. All the submissions have been peer reviewed from at least 3 independent
reviewers from our strong international program committee in order to ensure high
quality of contributed material as well as adherence to the conference topics. After
the review process, 65 papers were selected to be included in the proceedings.
Each edition of the workshop represents a challenge that cannot be successfully
achieved without the deep involvement of numerous people and institutions. We
address sincere thanks to all of them for their support, and to the University of Milan
for making us so welcome.
We are very grateful to our keynote speakers for their plenary lectures covering
different areas of the conference. The talk of Guido Caldarelli (IMT Lucca - Italy)
focused on the origins of instability in financial networks. The presentation given by
Raissa D’Souza (U. C. Davis - USA) dealt with the steering and controlling systems
of interdependent networks. Renaud Lambiotte (University of Namur -Belgium)
gave a talk on “Burstiness and spreading on networks: models and predictions” and
Yamir Moreno (University of Zaragoza - Spain) presented the talk “On the structure
and dynamics of multilayer networks”. The talk given by Eiko Yoneki (University
of Cambridge - UK) was about “Efficient large-scale graph processing” and Ben
Y. Zhao (U. C. Santa-Barbara - USA) covered the link prediction issue from an
empirical perspective. Their support of the workshop is without a doubt one of the
reasons of the success of COMPLEX NETWORKS 2016.
Two speakers gave very illuminating tutorials that drew many conference partici-
pants. These talks, held on November 29, 2016 were accessible to a general audience
of graduate students. Ernesto Estrada (University of Strathclyde Glasgow - UK)
gave a lecture on “Consensus dynamics on networks. Theory and applications” and
Bruno Gonçalves (New York University - USA) delivered a practical introduction to
machine learning (with Python).
We record our thanks to our fellow members of the Organizing Committee:
Chantal Cherifi (University of Lyon2 - France) and Antonio Scala (CNR - Italy), the
poster chairs, for arranging the poster session program and the editing of the book
of abstracts; Bruno Gonçalves (New York University - USA), the publicity chair,
for his work in securing a substantial input of papers from both Asia and America
and in encouraging participation from those areas; and all the session chairs for their
outstanding participation. We would also like to record our appreciation for the work
of the Local Arrangement Committee. In particular, Carlo Piccardi (Politecnico di
Milano - Italy) and Fabio Della Rossa (Politecnico di Milano - Italy) in making all
the excellent logistical arrangements for the conference. We also acknowledge the
important contributions of the members of the Computer Science Department of the
University of Milan. In particular, the team of the NPTLab ( University of Milan)
led by Gian Paolo Rossi. We thank him for his unwavering support. Many thanks
to its junior members, Matteo Zignani and Christan Quadri for the incredible work
they have done in the organization and the editing of the proceeding. We extend our
thanks to Matteo Re and Giorgio Valentini, their efforts made a great contribution to
the success of the workshop.
Preface vii
We are also indebted to our partners, Alessandro Fellegara and Alessandro Egro
along with their team (Tribe Communication) for their passion and patience in
designing the visual identity of the workshop. Our gratitude must also be extended to
our sponsors, Blogmeter, Celi and Shaman, for supporting the workshop.
We would also like to express our deepest appreciation to all those who have
helped us for the success of this meeting. Sincere thanks to the contributors, the
success of the technical program would not be possible without their creativity.
Finally, we would like to express our most sincere thanks to the Program Committee
members who have so generously volunteered their precious time to support the peer
review process.
We hope that this volume makes a useful contribution to issues surrounding the
fascinating world of complex networks and that you enjoy the papers as much as we
enjoyed organizing the conference and putting this collection of papers together.
ix
x Organization & Committees
xv
xvi Contents
Stock prices prediction via tensor decomposition and links forecast . . . . . . 645
Alessandro Spelta
Who buys what, where: Reconstruction of the international trade flows
by commodity and industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 657
Yuichi Ikeda and Tsutomu Watanabe
1 Introduction
Understanding edge formation in networks is a key interest of our research commu-
nity. For example, social scientists are frequently interested in studying relations
between entities within social networks, e.g., how social friendship ties form between
actors and explain them based on attributes such as a person’s gender, race, political
affiliation or age in the network [18]. Similarly, the complex networks community
suggests a set of generative network models aiming at explaining the formation of
edges focusing on the two core principles of popularity and similarity [15]. Thus,
a series of approaches to study edge formation have emerged including statistical
(regression) tools [10, 23] and model-based approaches [6, 15, 24] specifically estab-
lished in the physics and complex networks communities. Other disciplines such as
the computer sciences, biomedical sciences or political sciences use these tools to
!
! " ! " ! !
model and encode and integrate them as priors into the Bayesian framework. Using
Bayes factors with marginal likelihood estimations allows us to compare the relative
plausibility of expressed hypotheses as they are specifically sensitive to the priors.
The final output is a ranking of hypotheses based on their plausibility given the data.
Contributions. Our main contributions are: (i) We present a first step towards a
Bayesian approach for comparing generative hypotheses about edge formation in
networks. (ii) We provide simple categorical models based on local and global
scenarios allowing the comparison of hypotheses for multigraphs. (iii) We provide
guidelines for building hypotheses based on node attributes. (iv) We demonstrate
the applicability of our approach on synthetic and empirical data. (v) We make an
implementation of this approach openly available1 on the Web.
2 Background
We start by introducing the underlying concepts of our approach.
Attributed Multigraphs. In this paper, we focus on multigraphs with attributed
nodes and unweighted edges without own identity. That means, each pair of nodes
can be connected by multiple indistinguishable edges, and there are features for the
individual nodes available.
We formally define this as: Let G = (V, E, F) be an unweighted attributed multi-
graph with V = (v1 , . . . , vn ) being a list of nodes, E = {(vi , v j )} ∈ V ×V a multiset of
either directed or undirected edges, and a set of feature vectors F = ( f1 , . . . , fn ). Each
feature vector fi = ( fi [1], ..., fi [c])T maps a node vi to c (numeric or categorical) at-
tribute values. The graph structure is captured by an adjacency matrix Mn×n = (mi j ),
where mi j is the multiplicity of edge (vi , v j ) in E (i.e., number of edges between
nodes vi and v j ). By definition, the total number of multiedges is l = |E| = ∑i j mi j .
Fig. 1a shows an example unweighted attributed multigraph: nodes represent
authors, and undirected edges represent co-authorship in scientific articles. The
adjacency matrix of this graph—counting for multiplicity of edges—is shown in Fig.
1b. Feature vectors (node attributes) are described in Fig. 1c. Thus, for this particular
case, we account for n = 4 nodes, l = 44 multiedges, and c = 6 attributes.
Bayesian Hypothesis Testing. Our approach compares hypotheses on edge forma-
tion based on techniques from Bayesian hypothesis testing [11, 21]. The elementary
Bayes’ theorem states for parameters θ , given data D and a hypothesis H that:
likelihood prior
posterior z }| { z }| {
z }| { P(D|θ , H) P(θ |H)
P(θ |D, H) = (1)
P(D|H)
| {z }
marginal likelihood
As observed data D, we use the adjacency matrix M, which encodes edges counts. θ
refers to the model parameters, which in our scenario correspond to the probabilities
of individual edges. H denotes a hypothesis under investigation. The likelihood
1 https://github.com/lisette-espin/JANUS
6 Lisette Espı́n-Noboa,Florian Lemmerich, Markus Strohmaier and Philipp Singer
3 Approach
In this section, we describe the main steps towards a hypotheses-driven Bayesian
approach for understanding edge formation in unweighted attributed multigraphs. To
that end, we propose intuitive models for edge formation (Section 3.1), a flexible
toolbox to formally specify belief in the model parameters (Section 3.2), a way of
computing proper (Dirichlet) priors from these beliefs (Section 3.2), computation
of the marginal likelihood in this scenario (Section 3.3), and guidelines on how to
interpret the results (Section 3.4). We subsequently discuss these issues one-by-one.
n Γ (∑nj=1 αi j ) n Γ (αi j + mi j )
P(D|H) = ∏ n ∏ (2)
i=1 Γ (∑ j=1 αi j + mi j ) j=1 Γ (αi j )
Recall, αi j encodes our prior belief connecting nodes vi and v j in G, and mi j are
the actual edge counts. Since we evaluate only a single model in the global case, the
product over rows i of the adjacency matrix can be removed, and we obtain:
Γ (∑ni=1 ∑nj=1 αi j ) n n Γ (α + m )
ij ij
P(D|H) = (3)
Γ (∑ni=1 ∑nj=1 αi j + mi j ) ∏ ∏ Γ (αi j )
i=1 j=1
Equation (3) holds for directed networks. In the undirected case, indices j go
from i to n accounting for only half of the matrix including the diagonal. For a
detailed derivation of the marginal likelihood given a Dirichlet-Categorical model
see [22, 25]. For both models we focus on the log-marginal likelihoods in practice to
avoid underflows.
Bayes Factor. Formally, we compare the relative plausibility of hypotheses by using
so-called Bayes factors [7], which simply are the ratios of the marginal likelihoods
for two hypotheses H1 and H2 . If it is positive, the first hypothesis is judged as more
plausible. The strength of the Bayes factor can be checked in an interpretation table
provided by Kass and Raftery [7].
!
" "!
!
#
4 Experiments
We demonstrate the utility of our approach on both synthetic and empirical networks.
Due to space limitations, we only showcase the local model results.
Updated editions will replace the previous one—the old editions will
be renamed.
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the terms
of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.
• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
about donations to the Project Gutenberg Literary Archive
Foundation.”
• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.
1.F.
1.F.4. Except for the limited right of replacement or refund set forth in
paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.
Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.
Most people start at our website which has the main PG search
facility: www.gutenberg.org.