-
Detecting hierarchical and overlapping network communities using locally optimal modularity changes
Authors:
Michael J. Barber
Abstract:
Agglomerative clustering is a well established strategy for identifying communities in networks. Communities are successively merged into larger communities, coarsening a network of actors into a more manageable network of communities. The order in which merges should occur is not in general clear, necessitating heuristics for selecting pairs of communities to merge. We describe a hierarchical clu…
▽ More
Agglomerative clustering is a well established strategy for identifying communities in networks. Communities are successively merged into larger communities, coarsening a network of actors into a more manageable network of communities. The order in which merges should occur is not in general clear, necessitating heuristics for selecting pairs of communities to merge. We describe a hierarchical clustering algorithm based on a local optimality property. For each edge in the network, we associate the modularity change for merging the communities it links. For each community vertex, we call the preferred edge that edge for which the modularity change is maximal. When an edge is preferred by both vertices that it links, it appears to be the optimal choice from the local viewpoint. We use the locally optimal edges to define the algorithm: simultaneously merge all pairs of communities that are connected by locally optimal edges that would increase the modularity, redetermining the locally optimal edges after each step and continuing so long as the modularity can be further increased. We apply the algorithm to model and empirical networks, demonstrating that it can efficiently produce high-quality community solutions. We relate the performance and implementation details to the structure of the resulting community hierarchies. We additionally consider a complementary local clustering algorithm, describing how to identify overlapping communities based on the local optimality condition.
△ Less
Submitted 4 September, 2013; v1 submitted 27 May, 2013;
originally announced May 2013.
-
NetzCope: A Tool for Displaying and Analyzing Complex Networks
Authors:
Michael J. Barber,
Ludwig Streit,
Oleg Strogan
Abstract:
Networks are a natural and popular mechanism for the representation and investigation of a broad class of systems. But extracting information from a network can present significant challenges. We present NetzCope, a software application for the display and analysis of networks. Its key features include the visualization of networks in two or three dimensions, the organization of vertices to reveal…
▽ More
Networks are a natural and popular mechanism for the representation and investigation of a broad class of systems. But extracting information from a network can present significant challenges. We present NetzCope, a software application for the display and analysis of networks. Its key features include the visualization of networks in two or three dimensions, the organization of vertices to reveal structural similarity, and the detection and visualization of network communities by modularity maximization.
△ Less
Submitted 7 July, 2010; v1 submitted 22 June, 2010;
originally announced June 2010.
-
Noise and Neuronal Heterogeneity
Authors:
Michael J. Barber,
Manfred L. Ristig
Abstract:
We consider signal transaction in a simple neuronal model featuring intrinsic noise. The presence of noise limits the precision of neural responses and impacts the quality of neural signal transduction. We assess the signal transduction quality in relation to the level of noise, and show it to be maximized by a non-zero level of noise, analogous to the stochastic resonance effect. The quality enha…
▽ More
We consider signal transaction in a simple neuronal model featuring intrinsic noise. The presence of noise limits the precision of neural responses and impacts the quality of neural signal transduction. We assess the signal transduction quality in relation to the level of noise, and show it to be maximized by a non-zero level of noise, analogous to the stochastic resonance effect. The quality enhancement occurs for a finite range of stimuli to a single neuron; we show how to construct networks of neurons that extend the range. The range increases more rapidly with network size when we make use of heterogeneous populations of neurons with a variety of thresholds, rather than homogeneous populations of neurons all with the same threshold. The limited precision of neural responses thus can have a direct effect on the optimal network structure, with diverse functional properties of the constituent neurons supporting an economical information processing strategy that reduces the metabolic costs of handling a broad class of stimuli.
△ Less
Submitted 4 May, 2010;
originally announced May 2010.
-
Distinct spatial characteristics of industrial and public research collaborations: Evidence from the 5th EU Framework Programme
Authors:
Thomas Scherngell,
Michael J. Barber
Abstract:
This study compares the spatial characteristics of industrial R&D networks to those of public research R&D networks (i.e. universities and research organisations). The objective is to measure the impact of geographical separation effects on the constitution of cross-region R&D collaborations for both types of collaboration. We use data on joint research projects funded by the 5th European Framewor…
▽ More
This study compares the spatial characteristics of industrial R&D networks to those of public research R&D networks (i.e. universities and research organisations). The objective is to measure the impact of geographical separation effects on the constitution of cross-region R&D collaborations for both types of collaboration. We use data on joint research projects funded by the 5th European Framework Programme (FP) to proxy cross-region collaborative activities. The study area is composed of 255 NUTS-2 regions that cover the EU-25 member states (excluding Malta and Cyprus) as well as Norway and Switzerland. We adopt spatial interaction models to analyse how the variation of cross-region industry and public research networks is affected by geography. The results of the spatial analysis provide evidence that geographical factors significantly affect patterns of industrial R&D collaboration, while in the public research sector effects of geography are much smaller. However, the results show that technological distance is the most important factor for both industry and public research cooperative activities.
△ Less
Submitted 21 April, 2010;
originally announced April 2010.
-
Spatial Interaction Modelling of Cross-Region R&D Collaborations: Empirical Evidence from the 5th EU Framework Programme
Authors:
Thomas Scherngell,
Michael J. Barber
Abstract:
The last few years have witnessed an increasing interest in the geography of innovation. As noted by Autant-Bernard et al. (2007a), the geographical dimension of innovation deserves further attention by analysing such phenomena as R&D collaborations. In this study we focus on cross-region R&D collaborations in Europe. The European coverage is achieved by using data on collaborative R&D projects fu…
▽ More
The last few years have witnessed an increasing interest in the geography of innovation. As noted by Autant-Bernard et al. (2007a), the geographical dimension of innovation deserves further attention by analysing such phenomena as R&D collaborations. In this study we focus on cross-region R&D collaborations in Europe. The European coverage is achieved by using data on collaborative R&D projects funded by the EU Framework Programmes (FPs) between organisation that are located in 255 NUTS-2 regions of the 25 pre-2007 EU member-states, as well as Norway and Switzerland. The objective is to identify separation effects - such as geographical or technological effects - on the constitution of cross-region collaborative R&D activities. We specify a Poisson spatial interaction model to analyse these questions. The dependent variable is the intensity of cross-region R&D collaborations, the independent variables include origin, destination and separation characteristics of interaction. The results provide striking evidence that geographical effects play an important role in determining R&D collaborations across Europe. Geographical proximity and co-localisation of organisations in neighbouring regions are important determinants of cross-region collaboration intensities. The effect of technological proximity is stronger than spatial effects. R&D collaborations occur most often between organisations that are located close to each other in technological space. R&D collaborations in Europe are also affected by language- and country border effects, but these are smaller than geographical and technological effects.
△ Less
Submitted 20 April, 2010;
originally announced April 2010.
-
Analyzing and modeling European R&D collaborations: Challenges and opportunities from a large social network
Authors:
Michael J. Barber,
Manfred Paier,
Thomas Scherngell
Abstract:
Networks have attracted a burst of attention in the last decade, with applications to natural, social, and technological systems. While networks provide a powerful abstraction for investigating relationships and interactions, the preparation and analysis of complex real-world networks nonetheless presents significant challenges. In particular social networks are characterized by a large number of…
▽ More
Networks have attracted a burst of attention in the last decade, with applications to natural, social, and technological systems. While networks provide a powerful abstraction for investigating relationships and interactions, the preparation and analysis of complex real-world networks nonetheless presents significant challenges. In particular social networks are characterized by a large number of different properties and generation mechanisms which require a rich set of indicators. The objective of the current study is to analyze large social networks with respect to their community structure and mechanisms of network formation. As a case study, we consider networks derived from the European Union's Framework Programs for Research and Technological Development using methods from statistical physics and economic geography.
△ Less
Submitted 8 April, 2010;
originally announced April 2010.
-
Community Structure and Topical Differentiation in European RTD Collaborations
Authors:
Michael J. Barber,
Margarida Faria,
Ludwig Streit,
Oleg Strogan
Abstract:
We investigate research and development collaborations under the EU Framework Programs (FPs) for Research and Technological Development. The collaborations in the FPs give rise to bipartite networks, with edges existing between projects and the organizations taking part in them. A version of the modularity measure, adapted to bipartite networks, is presented. Communities are found so as to maximiz…
▽ More
We investigate research and development collaborations under the EU Framework Programs (FPs) for Research and Technological Development. The collaborations in the FPs give rise to bipartite networks, with edges existing between projects and the organizations taking part in them. A version of the modularity measure, adapted to bipartite networks, is presented. Communities are found so as to maximize the bipartite modularity. Projects in the resulting communities are shown to be topically differentiated.
△ Less
Submitted 31 March, 2010;
originally announced March 2010.
-
Investigating an online social network using spatial interaction models
Authors:
Conrad Lee,
Thomas Scherngell,
Michael J. Barber
Abstract:
Using a spatial interaction modeling approach, we investigate the German collegiate social network site StudiVZ. We focus on identifying factors that foster strong inter-institutional linkages, testing whether the acquaintanceship rate between institutions of higher education is related to various geographic and institutional attributes. We find that acquaintanceship is most significantly related…
▽ More
Using a spatial interaction modeling approach, we investigate the German collegiate social network site StudiVZ. We focus on identifying factors that foster strong inter-institutional linkages, testing whether the acquaintanceship rate between institutions of higher education is related to various geographic and institutional attributes. We find that acquaintanceship is most significantly related to geographic separation: measuring distance with automobile travel time, acquaintanceship drops by 91% for each additional 100 minutes. Institution type and the former East-West German divide are also related to this rate with statistical significance.
△ Less
Submitted 17 November, 2010; v1 submitted 6 November, 2009;
originally announced November 2009.
-
Comparing university organizational units and scientific co-authorship communities
Authors:
Uwe Obermeier,
Michael J. Barber,
Andreas Krueger,
Hannes Brauckmann
Abstract:
A co-authorship network of scientists at a university is an archetypical example of a complex evolving network. Collaborative R&D networks are self-organized products of partner choice between scientists. Modern science is, due to the immanent imperative of newness, strongly interdisciplinary. Crossovers between the different scientific disciplines and organizational units are observable on a da…
▽ More
A co-authorship network of scientists at a university is an archetypical example of a complex evolving network. Collaborative R&D networks are self-organized products of partner choice between scientists. Modern science is, due to the immanent imperative of newness, strongly interdisciplinary. Crossovers between the different scientific disciplines and organizational units are observable on a daily basis. Since collaborative research has become the dominant and most promising way to produce high-quality output, collaboration structures are also a target for research and management design.
We study co-authorship covered by the Citation Index within University College Dublin, a large Irish university. We focus especially on collaborations between organizational units, such as schools or colleges. We compare the centrality of the brokerage individuals within their organizational units. The network of co-authorship is analyzed with standard network measures, with typical features for complex networks observed. We use the bipartite network of authors and papers for the identification of communities of collaborators and clusters of papers. The observed communities are compared to the formal organizational units of the university.
△ Less
Submitted 31 August, 2009;
originally announced August 2009.
-
Detecting network communities by propagating labels under constraints
Authors:
Michael J. Barber,
John W. Clark
Abstract:
We investigate the recently proposed label-propagation algorithm (LPA) for identifying network communities. We reformulate the LPA as an equivalent optimization problem, giving an objective function whose maxima correspond to community solutions. By considering properties of the objective function, we identify conceptual and practical drawbacks of the label propagation approach, most importantly…
▽ More
We investigate the recently proposed label-propagation algorithm (LPA) for identifying network communities. We reformulate the LPA as an equivalent optimization problem, giving an objective function whose maxima correspond to community solutions. By considering properties of the objective function, we identify conceptual and practical drawbacks of the label propagation approach, most importantly the disparity between increasing the value of the objective function and improving the quality of communities found. To address the drawbacks, we modify the objective function in the optimization problem, producing a variety of algorithms that propagate labels subject to constraints; of particular interest is a variant that maximizes the modularity measure of community quality. Performance properties and implementation details of the proposed algorithms are discussed. Bipartite as well as unipartite networks are considered.
△ Less
Submitted 18 June, 2009; v1 submitted 18 March, 2009;
originally announced March 2009.
-
Searching for Communities in Bipartite Networks
Authors:
Michael J. Barber,
Margarida Faria,
Ludwig Streit,
Oleg Strogan
Abstract:
Bipartite networks are a useful tool for representing and investigating interaction networks. We consider methods for identifying communities in bipartite networks. Intuitive notions of network community groups are made explicit using Newman's modularity measure. A specialized version of the modularity, adapted to be appropriate for bipartite networks, is presented; a corresponding algorithm is…
▽ More
Bipartite networks are a useful tool for representing and investigating interaction networks. We consider methods for identifying communities in bipartite networks. Intuitive notions of network community groups are made explicit using Newman's modularity measure. A specialized version of the modularity, adapted to be appropriate for bipartite networks, is presented; a corresponding algorithm is described for identifying community groups through maximizing this measure. The algorithm is applied to networks derived from the EU Framework Programs on Research and Technological Development. Community groups identified are compared using information-theoretic methods.
△ Less
Submitted 19 March, 2008;
originally announced March 2008.
-
Modularity and community detection in bipartite networks
Authors:
Michael J. Barber
Abstract:
The modularity of a network quantifies the extent, relative to a null model network, to which vertices cluster into community groups. We define a null model appropriate for bipartite networks, and use it to define a bipartite modularity. The bipartite modularity is presented in terms of a modularity matrix B; some key properties of the eigenspectrum of B are identified and used to describe an al…
▽ More
The modularity of a network quantifies the extent, relative to a null model network, to which vertices cluster into community groups. We define a null model appropriate for bipartite networks, and use it to define a bipartite modularity. The bipartite modularity is presented in terms of a modularity matrix B; some key properties of the eigenspectrum of B are identified and used to describe an algorithm for identifying modules in bipartite networks. The algorithm is based on the idea that the modules in the two parts of the network are dependent, with each part mutually being used to induce the vertices for the other part into the modules. We apply the algorithm to real-world network data, showing that the algorithm successfully identifies the modular structure of bipartite networks.
△ Less
Submitted 5 November, 2007; v1 submitted 11 July, 2007;
originally announced July 2007.
-
Expectation-driven interaction: a model based on Luhmann's contingency approach
Authors:
M. J. Barber,
Ph. Blanchard,
E. Buchinger,
B. Cessac,
L. Streit
Abstract:
We introduce an agent-based model of interaction, drawing on the contingency approach from Luhmann's theory of social systems. The agent interactions are defined by the exchange of distinct messages. Message selection is based on the history of the interaction and developed within the confines of the problem of double contingency. We examine interaction strategies in the light of the message-exc…
▽ More
We introduce an agent-based model of interaction, drawing on the contingency approach from Luhmann's theory of social systems. The agent interactions are defined by the exchange of distinct messages. Message selection is based on the history of the interaction and developed within the confines of the problem of double contingency. We examine interaction strategies in the light of the message-exchange description using analytical and computational methods.
△ Less
Submitted 19 December, 2006;
originally announced December 2006.
-
Multiple Thresholds in a Model System of Noisy Ion Channels
Authors:
Michael J. Barber,
Manfred L. Ristig
Abstract:
Voltage-activated ion channels vary randomly between open and closed states, influenced by the membrane potential and other factors. Signal transduction is enhanced by noise in a simple ion channel model. The enhancement occurs in a finite range of signals; the range can be extended using populations of channels. The range increases more rapidly in multiple-threshold channel populations than in…
▽ More
Voltage-activated ion channels vary randomly between open and closed states, influenced by the membrane potential and other factors. Signal transduction is enhanced by noise in a simple ion channel model. The enhancement occurs in a finite range of signals; the range can be extended using populations of channels. The range increases more rapidly in multiple-threshold channel populations than in single-threshold populations. The diversity of ion channels may thus be present as a strategy to reduce the metabolic costs of handling a broad class of electrochemical signals.
△ Less
Submitted 1 September, 2006; v1 submitted 23 February, 2006;
originally announced February 2006.
-
The Network of EU-Funded Collaborative R&D Projects
Authors:
M. J. Barber,
A. Krueger,
T. Krueger,
T. Roediger-Schluga
Abstract:
We describe collaboration networks consisting of research projects funded by the European Union and the organizations involved in those projects. The networks are of substantial size and complexity, but are important to understand due to the significant impact they could have on research policies and national economies in the EU. In empirical determinations of the network properties, we observe…
▽ More
We describe collaboration networks consisting of research projects funded by the European Union and the organizations involved in those projects. The networks are of substantial size and complexity, but are important to understand due to the significant impact they could have on research policies and national economies in the EU. In empirical determinations of the network properties, we observe characteristics similar to other collaboration networks, including scale-free degree distributions, small diameter, and high clustering. We present some plausible models for the formation and structure of networks with the observed properties.
△ Less
Submitted 20 October, 2005; v1 submitted 14 September, 2005;
originally announced September 2005.
