We present an analysis of a person-to-person recommendation network, consisting 4 million people who made 16 recommendations on half products. observe the propagation and cascade sizes, which we explain by simple stochastic model. analyze how user behavior varies within communities defined network. Product purchases follow ‘long tail’ where significant share belongs to rarely sold items. establish network grows over time effective it is from viewpoint sender receiver recommendations. While...

Many communication and social networks have power-law link distributions, containing a few nodes that very high degree many with low degree. The connectivity play the important role of hubs in networking, fact can be exploited when designing efficient search algorithms. We introduce number local strategies utilize graphs costs scaling sublinearly size graph. also demonstrate utility these on GNUTELLA peer-to-peer network.

Barabasi and Albert ([1][1]) propose an improved version of the Erdos-Renyi (ER) theory random networks to account for scaling properties a number systems, including link structure World Wide Web (WWW). The they present, however, is inconsistent with empirically

Early patterns of Digg diggs and YouTube views reflect long-term user interest.

Scholars, advertisers and political activists see massive online social networks as a representation of interactions that can be used to study the propagation ideas, bond dynamics viral marketing, among others. But linked structures do not reveal actual people. Scarcity attention daily rythms life work makes people default interacting with those few matter reciprocate their attention. A within Twitter reveals driver usage is sparse hidden network connections underlying “declared” set friends...

Collaborative tagging describes the process by which many users add metadata in form of keywords to shared content. Recently, collaborative has grown popularity on web, sites that allow tag bookmarks, photographs and other In this paper we analyze structure systems as well their dynamical aspects. Specifically, discovered regularities user activity, frequencies, kinds tags used, bursts bookmarking a remarkable stability relative proportions within given url. We also present model predicts...

We investigate the effectiveness of connectionist architectures for predicting future behavior nonlinear dynamical systems. focus on real-world time series limited record length. Two examples are analyzed: benchmark sunspot and chaotic data from a computational ecosystem. The problem overfitting, particularly serious short records noisy data, is addressed both by using statistical method validation adding complexity term to cost function ("back-propagation with weight-elimination")....

One of the most common modes accessing information in World Wide Web is surfing from one document to another along hyperlinks. Several large empirical studies have revealed patterns behavior. A model that assumes users make a sequence decisions proceed page, continuing as long value current page exceeds some threshold, yields probability distribution for number pages user visits within given site. This was verified by comparing its predictions with detailed measurements patterns. The also...

The authors have designed and implemented an open, market-based computational system called Spawn. Spawn utilizes idle resources in a distributed network of heterogeneous computer workstations. It supports both coarse-grain concurrent applications the remote execution many independent tasks. Using Monte Carlo simulations as prototypical applications, explore issues fairness resource distribution, currency form priority, price equilibria, dynamics transients, scaling to large systems. In...

The subject of collective attention is central to an information age where millions people are inundated with daily messages. It thus interest understand how novel items propagates and eventually fades among large populations. We have analyzed the dynamics 1 million users interactive web site, digg.com, devoted thousands news stories. observations can be described by a dynamical model characterized single novelty factor. Our measurements indicate that within groups decays...

The Prisoner's Dilemma has long been considered the paradigm for studying emergence of cooperation among selfish individuals. Because its importance, it studied through computer experiments as well in laboratory and by analytical means. However, there are important differences between way a system composed many interacting elements is simulated digital machine manner which behaves when real experiments. In some instances, these disparities can be marked enough so to cast doubt on...

We present an analysis of a person-to-person recommendation network, consisting 4 million people who made 16 recommendations on half products. observe the propagation and cascade sizes, which we explain by simple stochastic model. then establish how network grows over time effective it is from viewpoint sender receiver recommendations. While average are not very at inducing purchases do spread far, model that successfully identifies product pricing categories for viral marketing seems to be...

A general method for combining existing algorithms into new programs that are unequivocally preferable to any of the component is presented. This method, based on notions risk in economics, offers a computational portfolio design procedure can be used wide range problems. Tested by solving canonical NP-complete problem, problems ranging from combinatorics DNA sequencing completion tasks environments with resource contention, such as World Wide Web.

Scholars, advertisers and political activists see massive online social networks as a representation of interactions that can be used to study the propagation ideas, bond dynamics viral marketing, among others. But linked structures do not reveal actual people. Scarcity attention daily rythms life work makes people default interacting with those few matter reciprocate their attention. A within Twitter reveals driver usage is sparse hidden network connections underlying "declared" set friends...

The aperiodic or chaotic behavior for one-dimensional maps just before a tangent bifurcation occurs appears as intermittency in which long laminarlike regions irregularly separated by bursts occur. Proceeding from the picture proposed Pomeau and Manneville, numerical experiments analytic calculations are carried out on various models exhibiting this behavior. presence of external noise is analyzed, case general power dependence curve near studied. Scaling relations average length laminar...

We describe a method for the automatic identification of communities practice from e-mail logs within an organization. use betweenness centrality algorithm that can rapidly find graph representing information flows. apply this to initial corpus nearly 1 million messages collected over 2-month span, and show is effective at identifying true communities, both formal informal, these scale-free graphs. This approach also enables leadership roles communities. These studies are complemented by...

The ever-increasing amount of information flowing through Social Media forces the members these networks to compete for attention and influence by relying on other people spread their message. A large study propagation within Twitter reveals that majority users act as passive consumers do not forward content network. Therefore, in order individuals become influential they must only obtain thus be popular, but also overcome user passivity. We propose an algorithm determines passivity based...