Abstract: The concept of microcontent poses a new set of challenges for the design of recommender systems that can assist the users to accomplish a broad set of complex informational tasks as well as to evaluate the importance of information resources such as microcontent structures. In this paper we formulate an approach by using an adaptation of the hubs and authorities model, in order to study the deployment of a recommender system for a microcontent structure such as Wikipedia. 1.

The Eigenfactor score is a journal influence metric developed at the Department of Biology of the University of Washington and recently introduced in the Science and Social Science Journal Citation Reports maintained by Thomson Reuters. It provides a compelling measure of journal status with solid mathematical background, sound axiomatic foundation, intriguing stochastic interpretation, and many interesting relationships to other ranking measures. In this short contribution, we give ten reasons to motivate the use of the Eigenfactor method.

The difference between popularity and prestige in the sciences and in the social sciences: a bibliometric analysis ✩

Recently, some researchers have attempted to exploit state-aggregation techniques to compute stable distributions of high-dimensional Markov matrices (Gambin and Pokarowski, 2001). While these researchers have devised an efficient, recursive algorithm, their results are only approximate. We improve upon past results by presenting a novel state aggregation technique, which we use to give the first (to our knowledge) scalable, exact algorithm for computing the stochastically stable distribution of a perturbed Markov matrix. Since it is not combinatorial in nature, our algorithm is computationally feasible even for highdimensional

Node centrality measures are important in a large number of graph applications, from search and ranking to social and biological network analysis. In this paper we study node centrality for very large graphs, up to billions of nodes and edges. Various definitions for centrality have been proposed, ranging from very simple (e.g., node degree) to more elaborate. However, measuring centrality in billion-scale graphs poses several challenges. Many of the “traditional ” definitions such as closeness and betweenness were not designed with scalability in mind. Therefore, it is very difficult, if not impossible, to compute them both accurately and efficiently. In this paper, we propose centrality measures suitable for very large graphs, as well as scalable methods to effectively compute them. More specifically, we propose effective closeness and LINERANK which are designed for billion-scale graphs. We also develop algorithms to compute the proposed centrality measures in MAPREDUCE, a modern paradigm for large-scale, distributed data processing. We present extensive experimental results on both synthetic and real datasets, which demonstrate the scalability of our approach to very large graphs, as well as interesting findings and anomalies. 1

In this paper, we introduce the notion of derivatives of centrality metrics for graph visualizations. As centrality represents the prestige or importance of a node in a network, its derivative with respect to any other node represents the influencing power it has over that node. Therefore, derivatives tell us how much a given node influences the importance of another node, even if they are not directly connected. We study three different centrality metrics and show the different results when visualizing their derivatives for a number of social and other scale-free networks. We show that derivatives not only provide an analysis tool for social networks, and also help us simplify the layout of complex networks in a way that retains the main structural properties. Centrality derivatives also help to visually measure the centralization degree of a network and provide the necessary information for estimating other metrics, such as structural balance and uncertainty. Through a number of examples, we show the flexibility and generality of this approach, and a general mechanism for extending this to any centrality metric.

In a previous paper, we have found empirical evidence supporting a positive relationship between network centrality and success. However, we have also found that more successful projects have a lower technical quality. A first, straightforward argument explaining previous findings is that more central contributors are also highly skilled developers who are well known for their ability to manage the complexity of code with a lower attention to the software structure. The consolidated metrics of software quality used by the authors in their previous research represent measures of code structure. This paper provides empirical evidence supporting the idea that the negative impact of success on quality is caused by the careless behaviour of skilled developers, who are also hubs within the social network. Research hypotheses are tested on a sample of 56 OS applications from the SourceForge.net repository, with a total of 378 developers. The sample includes some of the most successful and large OS projects, as well as a cross-section of less famous active projects evenly distributed among SourceForge.net’s project categories.

contained in this dissertation are those of the author and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the

Abstract. We explore a simple mathematical model of network computation, based on Markov chains. Similar models apply to a broad range of computational phenomena, arising in networks of computers, as well as in genetic, and neural nets, in social networks, and so on. The main problem of interaction with such spontaneously evolving computational systems is that the data are not uniformly structured. An interesting approach is to try to extract the semantical content of the data from their distribution among the nodes. A concept is then identified by finding the community of nodes that share it. The task of data structuring is thus reduced to the task of finding the network communities, as groups of nodes that together perform some non-local data processing. Towards this goal, we extend the ranking methods from nodes to paths, which allows us to extract information about the likely flow biases from the available static information about the network. 1

Abstract. Bibliometrics has the ambitious goal of measuring science. To this end, it exploits the way science is disseminated trough scientific publications and the resulting citation network of scientific papers. We survey the main historical contributions to the field, the most interesting bibliometric indicators, and the most popular bibliometric data sources. Moreover, we discuss distributions commonly used to model bibliometric phenomena and give an overview of methods to build bibliometric maps of science. 1