Andrea Capocci

Preferential attachment in the growth of social networks: The internet encyclopedia Wikipedia

We present an analysis of the statistical properties and growth of the free on-line encyclopedia Wikipedia. By describing topics by vertices and hyperlinks between them as edges, we can represent this encyclopedia as a directed graph. The topological properties of this graph are in close analogy with those of the World Wide Web, despite the very different growth mechanism. In particular, we measure a scale-invariant distribution of the in and out degree and we are able to reproduce these features by means of a simple statistical model. As a major consequence, Wikipedia growth can be described by local rules such as the preferential attachment mechanism, though users, who are responsible of its evolution, can act globally on the network.

Detecting communities in large networks

We develop an algorithm to detect community structure in complex networks. The algorithm is based on spectral methods and takes into account weights and link orientation. Since the method detects efficiently clustered nodes in large networks even when these are not sharply partitioned, it turns to be specially suitable for the analysis of social and information networks. We test the algorithm on a large-scale data-set from a psychological experiment of word association. In this case, it proves to be successful both in clustering words, and in uncovering mental association patterns.

Self-organized network evolution coupled to extremal dynamics

The interplay between topology and dynamics in complex networks is a fundamental but widely unexplored problem. Here, we study this phenomenon on a prototype model in which the network is shaped by a dynamical variable. We couple the dynamics of the Bak–Sneppen evolution model with the rules of the so-called fitness network model for establishing the topology of a network; each vertex is assigned a ‘fitness’, and the vertex with minimum fitness and its neighbours are updated in each iteration. At the same time, the links between the updated vertices and all other vertices are drawn anew with a fitness-dependent connection probability. We show analytically and numerically that the system self-organizes to a non-trivial state that differs from what is obtained when the two processes are decoupled. A power-law decay of dynamical and topological quantities above a threshold emerges spontaneously, as well as a feedback between different dynamical regimes and the underlying correlation and percolation properties of the network.

Number of loops of size h in growing scale-free networks.

The hierarchical structure of scale-free networks has been investigated focusing on the scaling of the number N(h)(t) of loops of size h as a function of the system size. In particular, we have found the analytic expression for the scaling of N(h)(t) in the Barabási-Albert (BA) scale-free network. We have performed numerical simulations on the scaling law for N(h)(t) in the BA network and in other growing scale-free networks, such as the bosonic network and the aging nodes network. We show that in the bosonic network and in the aging node network the phase transitions in the topology of the network are accompained by a change in the scaling of the number of loops with the system size.

Folksonomies and clustering in the collaborative system CiteULike

We analyze CiteULike, an online collaborative tagging system where users bookmark and annotate scientific papers. Such a system can be naturally represented as a tri-partite graph whose nodes represent papers, users and tags connected by individual tag assignments. The semantics of tags is studied here, in order to uncover the hidden relationships between tags. We find that the clustering coefficient can be used to analyze the semantical patterns among tags.

Taxonomy and clustering in collaborative systems: the case of the on-line encyclopedia Wikipedia

In this paper we investigate the nature and structure of the relation between imposed classifications and real clustering in a particular case of a scale-free network given by the on-line encyclopedia Wikipedia. We find a statistical similarity in the distributions of community sizes both by using the top-down approach of the categories division present in the archive and in the bottom-up procedure of community detection given by an algorithm based on the spectral properties of the graph. Regardless of the statistically similar behaviour, the two methods provide a rather different division of the articles, thereby signaling that the nature and presence of power laws is a general feature for these systems and cannot be used as a benchmark to evaluate the suitability of a clustering method.

Spectral methods cluster words of the same class in a syntactic dependency network

We analyze here a particular kind of linguistic network where vertices represent words and edges stand for syntactic relationships between words. The statistical properties of these networks have been recently studied and various features such as the small-world phenomenon and a scale-free distribution of degrees have been found. Our work focuses on four classes of words: verbs, nouns, adverbs and adjectives. Here, we use spectral methods sorting vertices. We show that the ordering clusters words of the same class. For nouns and verbs, the cluster size distribution clearly follows a power-law distribution that cannot be explained by a null hypothesis. Long-range correlations are found between vertices in the ordering provided by the spectral method. The findings support the use of spectral methods for detecting community structure.

Communities Detection in Large Networks

We develop an algorithm to detect community structure in complex networks. The algorithm is based on spectral methods and takes into account weights and links orientations. Since the method detects efficiently clustered nodes in large networks even when these are not sharply partitioned, it turns to be specially suitable to the analysis of social and information networks. We test the algorithm on a large-scale data-set from a psychological experiment of word association. In this case, it proves to be successful both in clustering words, and in uncovering mental association patterns.

Loops structure of the Internet at the autonomous system level.

We present here a study of the clustering and loops in a graph of the Internet at the autonomous systems level. We show that, even if the whole structure is changing with time, the statistical distributions of loops of order 3, 4, and 5 remain stable during the evolution. Moreover, we will bring evidence that the Internet graphs show characteristic Markovian signatures, since the structure is very well described by two-point correlations between the degrees of the vertices. This indeed proves that the Internet belongs to a class of network in which the two-point correlation is sufficient to describe their whole local (and thus global) structure. Data are also compared to present Internet models.

Quantitative description and modeling of real networks

We present data analysis and modeling of two particular cases of study in the field of growing networks. We analyze World Wide Web data set and authorship collaboration networks in order to check the presence of correlation in the data. The results are reproduced with good agreement through a suitable modification of the standard Albert-Barabási model of network growth. In particular, intrinsic relevance of sites plays a role in determining the future degree of the vertex.