Algorithms For Community Identification In Complex Networks

Complex networks such as the Internet, the World Wide Web (WWW), and various social and biological networks, are viewed as large, dynamic, random graphs, with properties significantly different from those of the Erdös-Rényi random graphs. In particular, properties such as degree distribution, network distance, transitivity and clustering coefficient of these networks have been empirically shown to diverge from classical random networks. Existence of communities is one such property inherent to these networks. A community may informally be defined as a locally-dense induced subgraph, of significant size, in a large globally-sparse graph. Recent empirical results reveal communities in networks spanning across different disciplines--physics, statistics, sociology, biology, and linguistics. At least two different questions may be posed on the community structure in large networks: (i) Given a network, detect or extract all (i.e., sets of nodes that constitute) communities; and (ii) Given a node in the network, identify the best community that the given node belongs to, if there exists one. Several algorithms have been proposed to solve the former problem, known as Community Discovery. The latter problem, known as Community Identification, has also been studied, but to a much smaller extent. Both these problems have been shown to be NP-complete, and a number of approximate algorithms have been proposed in recent years. A comprehensive taxonomy of the existing community detection algorithms is presented in this work. Global exploration of these complex networks to pull out communities (community discovery) is time and memory consuming. A more confined approach to mine communities in a given network is investigated in this research. Identifying communities does not require the knowledge of the entire graph. Community identification algorithms exist in the literature, but to a smaller extent. The dissertation presents a thorough description and analysis of the existing techniques to identify communities in large networks. Also a novel heuristic for identifying the community to which a given seed node belongs using only its neighborhood information is presented. An improved definition of a community based on the average degree of the induced subgraph is discussed thoroughly and it is compared with the various definitions in the literature. Next, a faster and accurate algorithm to identify communities in complex networks based on maximizing the average degree is described. The divisive nature of the algorithm (as against the existing agglomerative methods) efficiently identifies communities in large complex networks. The performance of the algorithm on several synthetic and real-world complex networks has also been thoroughly investigated.

This book constitutes the refereed proceedings of the twelfth International Conference, on Discovery Science, DS 2009, held in Porto, Portugal, in October 2009. The 35 revised full papers presented were carefully selected from 92 papers. The scope of the conference includes the development and analysis of methods for automatic scientific knowledge discovery, machine learning, intelligent data analysis, theory of learning, as well as their applications.

This book constitutes the refereed proceedings of the 10th International Conference on Parallel Problem Solving from Nature, PPSN 2008, held in Dortmund, Germany, in September 2008. The 114 revised full papers presented were carefully reviewed and selected from 206 submissions. The conference covers a wide range of topics, such as evolutionary computation, quantum computation, molecular computation, neural computation, artificial life, swarm intelligence, artificial ant systems, artificial immune systems, self-organizing systems, emergent behaviors, and applications to real-world problems. The paper are organized in topical sections on formal theory, new techniques, experimental analysis, multiobjective optimization, hybrid methods, and applications.

Identifying community structure can be a potent tool in the analysis and understanding of the structure of complex networks. Up to now, methods for evaluating the performance of identification algorithms use ad-hoc networks with communities of equal size. We show that inhomogeneities in community sizes can and do affect the performance of algorithms considerably, and propose an alternative method which takes these factors into account. Furthermore, we propose a simple modification of the algorithm proposed by Newman for community detection (Phys. Rev. E 69 066133) which treats communities of different sizes on an equal footing, and show that it outperforms the original algorithm while retaining its speed.

Graph partitioning and graph clustering are ubiquitous subtasks in many applications where graphs play an important role. Generally speaking, both techniques aim at the identification of vertex subsets with many internal and few external edges. To name only a few, problems addressed by graph partitioning and graph clustering algorithms are: What are the communities within an (online) social network? How do I speed up a numerical simulation by mapping it efficiently onto a parallel computer? How must components be organized on a computer chip such that they can communicate efficiently with each other? What are the segments of a digital image? Which functions are certain genes (most likely) responsible for? The 10th DIMACS Implementation Challenge Workshop was devoted to determining realistic performance of algorithms where worst case analysis is overly pessimistic and probabilistic models are too unrealistic. Articles in the volume describe and analyze various experimental data with the goal of getting insight into realistic algorithm performance in situations where analysis fails.

This book constitutes the refereed proceedings of the joint conference on Machine Learning and Knowledge Discovery in Databases: ECML PKDD 2008, held in Antwerp, Belgium, in September 2008. The 100 papers presented in two volumes, together with 5 invited talks, were carefully reviewed and selected from 521 submissions. In addition to the regular papers the volume contains 14 abstracts of papers appearing in full version in the Machine Learning Journal and the Knowledge Discovery and Databases Journal of Springer. The conference intends to provide an international forum for the discussion of the latest high quality research results in all areas related to machine learning and knowledge discovery in databases. The topics addressed are application of machine learning and data mining methods to real-world problems, particularly exploratory research that describes novel learning and mining tasks and applications requiring non-standard techniques.

The past decade has witnessed the emergence of participatory Web and social media, bringing people together in many creative ways. Millions of users are playing, tagging, working, and socializing online, demonstrating new forms of collaboration, communication, and intelligence that were hardly imaginable just a short time ago. Social media also helps reshape business models, sway opinions and emotions, and opens up numerous possibilities to study human interaction and collective behavior in an unparalleled scale. This lecture, from a data mining perspective, introduces characteristics of social media, reviews representative tasks of computing with social media, and illustrates associated challenges. It introduces basic concepts, presents state-of-the-art algorithms with easy-to-understand examples, and recommends effective evaluation methods. In particular, we discuss graph-based community detection techniques and many important extensions that handle dynamic, heterogeneous networks in social media. We also demonstrate how discovered patterns of communities can be used for social media mining. The concepts, algorithms, and methods presented in this lecture can help harness the power of social media and support building socially-intelligent systems. This book is an accessible introduction to the study of \emph{community detection and mining in social media}. It is an essential reading for students, researchers, and practitioners in disciplines and applications where social media is a key source of data that piques our curiosity to understand, manage, innovate, and excel. This book is supported by additional materials, including lecture slides, the complete set of figures, key references, some toy data sets used in the book, and the source code of representative algorithms. The readers are encouraged to visit the book website for the latest information. Table of Contents: Social Media and Social Computing / Nodes, Ties, and Influence / Community Detection and Evaluation / Communities in Heterogeneous Networks / Social Media Mining

An in-depth account of graph theory, written for serious students of mathematics and computer science. It reflects the current state of the subject and emphasises connections with other branches of pure mathematics. Recognising that graph theory is one of several courses competing for the attention of a student, the book contains extensive descriptive passages designed to convey the flavour of the subject and to arouse interest. In addition to a modern treatment of the classical areas of graph theory, the book presents a detailed account of newer topics, including Szemerédis Regularity Lemma and its use, Shelahs extension of the Hales-Jewett Theorem, the precise nature of the phase transition in a random graph process, the connection between electrical networks and random walks on graphs, and the Tutte polynomial and its cousins in knot theory. Moreover, the book contains over 600 well thought-out exercises: although some are straightforward, most are substantial, and some will stretch even the most able reader.

This book highlights cutting-edge research in the field of network science, offering scientists, researchers, students and practitioners a unique update on the latest advances in theory and a multitude of applications. It presents the peer-reviewed proceedings of the VI International Conference on Complex Networks and their Applications (COMPLEX NETWORKS 2017), which took place in Lyon on November 29 – December 1, 2017. The carefully selected papers cover a wide range of theoretical topics such as network models and measures; community structure, network dynamics; diffusion, epidemics and spreading processes; resilience and control as well as all the main network applications, including social and political networks; networks in finance and economics; biological and ecological networks and technological networks.

Community structure is a salient structural characteristic of many real-world networks. Communities are generally hierarchical, overlapping, multi-scale and coexist with other types of structural regularities of networks. This poses major challenges for conventional methods of community detection. This book will comprehensively introduce the latest advances in community detection, especially the detection of overlapping and hierarchical community structures, the detection of multi-scale communities in heterogeneous networks, and the exploration of multiple types of structural regularities. These advances have been successfully applied to analyze large-scale online social networks, such as Facebook and Twitter. This book provides readers a convenient way to grasp the cutting edge of community detection in complex networks. The thesis on which this book is based was honored with the “Top 100 Excellent Doctoral Dissertations Award” from the Chinese Academy of Sciences and was nominated as the “Outstanding Doctoral Dissertation” by the Chinese Computer Federation.