Human behavior in online social systems

We present and study data concerning human behavior in four online social systems: (i) an Internet community of friends of over 10⁷ people, (ii) a music community website with over 10⁶ users, (iii) a gamers’ community server with over 5 × 10⁶ users and (iv) a booklovers’ website with over 2.5 × 10⁵ users. The purpose of those systems is different; however, their properties are very similar. We have found that the distribution of human activity (e.g., the sum of books read or songs played) has the form of a power law. Moreover, the relationship between human activity and time has a power-law form, too. We present a simple interest-driven model of the evolution of such systems which explains the emergence of two scaling regimes.     

Hierarchy,cities size distribution and Zipf's law

We show that a hierarchical cities structure can be generated by a self-organized process which grows with a bottom-up mechanism, and that the resulting distribution is power law. First we analytically prove that the power law distribution satisfies the balance between the offer of the city and the demand of its basin of attraction, and that the exponent in the Zipf's law corresponds to the multiplier linking the population of the central city to the population of its basin of attraction. Moreover, the corresponding hierarchical structure shows a variable spanning factor, and the population of the cities linked to the same city up in the hierarchy is variable as well. Second a stochastic dynamic spatial model is proposed, whose numerical results confirm the analytical findings. In this model, inhabitants minimize the transportation cost, so that the greater the importance of this cost, the more stable is the system in its microscopic aspect. After a comparison with the existent methods for the generation of a power law distribution, conclusions are drawn on the connection of hierarchical structure, and power law distribution, with the functioning of the system of cities.     

Scaling and allometry in the building geometries of Greater London

Many aggregate distributions of urban activities such as city sizes reveal scaling but hardly any work exists on the properties of spatial distributions within individual cities, notwithstanding considerable knowledge about their fractal structure. We redress this here by examining scaling relationships in a world city using data on the geometric properties of individual buildings. We first summarise how power laws can be used to approximate the size distributions of buildings, in analogy to city-size distributions which have been widely studied as rank-size and lognormal distributions following Zipf [Human Behavior and the Principle of Least Effort (Addison-Wesley, Cambridge, 1949)] and Gibrat [Les Inégalités économiques (Librarie du Recueil Sirey, Paris, 1931)]. We then extend this analysis to allometric relationships between buildings in terms of their different geometric size properties. We present some preliminary analysis of building heights from the Emporis database which suggests very strong scaling in world cities. The data base for Greater London is then introduced from which we extract 3.6 million buildings whose scaling properties we explore. We examine key allometric relationships between these different properties illustrating how building shape changes according to size, and we extend this analysis to the classification of buildings according to land use types. We conclude with an analysis of two-point correlation functions of building geometries which supports our non-spatial analysis of scaling.     

Why are large cities faster? Universal scaling and self-similarity in urban organization and dynamics

Cities have existed since the beginning of civilization and have always been intimately connected with humanity's cultural and technological development. Much about the human and social dynamics that takes place is cities is intuitively recognizable across time, space and culture; yet we still do not have a clear cut answer as to why cities exist or to what factors are critical to make them thrive or collapse. Here, we construct an extensive quantitative characterization of the variation of many urban indicators with city size, using large data sets for American, European and Chinese cities. We show that social and economic quantities, characterizing the creation of wealth and new ideas, show increasing returns to population scale, which appear quantitatively as a power law of city size with an exponent β≃ 1.15 > 1. Concurrently, quantities characterizing material infrastructure typically show economies of scale, namely β≃ 0.8 < 1. The existence of pervasive scaling relations across city size suggests a universal social dynamics common to all cities within an urban system. We sketch some of their general ingredients, which include the acceleration of social life and a restructuring of individual social networks as cities grow larger. We also build simple dynamical models to show that increasing returns in wealth and innovation can fuel faster than exponential growth, which inexorably lead to crises of urban organization. To avoid them we show that growth may proceed in cycles, separated by major urban adaptations, with the unintended consequence that the duration of such cycles decreases with larger urban population size and is now estimated to be shorter than a human lifetime.     

Analysis of the Karmarkar-Karp differencing algorithm

The Karmarkar-Karp differencing algorithm is the best known polynomial time heuristic for the number partitioning problem, fundamental in both theoretical computer science and statistical physics. We analyze the performance of the differencing algorithm on random instances by mapping it to a nonlinear rate equation. Our analysis reveals strong finite size effects that explain why the precise asymptotics of the differencing solution is hard to establish by simulations. The asymptotic series emerging from the rate equation satisfies all known bounds on the Karmarkar-Karp algorithm and projects a scaling n ^{−c ln n} , where c = 1/(2 ln 2) = 0.7213 .... Our calculations reveal subtle relations between the algorithm and Fibonacci-like sequences, and we establish an explicit identity to that effect.     

An evolving network model in polymer melts with community structures

In order to describe the entangled network structure in polymer melts visually, we propose an evolving network model with community structure. This network model grows according to the inner-community and inter-community preferential mechanisms of both community sizes and node degrees. Numerical simulation results indicate that the cumulative distribution of community size and node degree distribution follow power-law distributions P(S≥s)∼s^-υ and P(k)∼k^-γ respectively, with the exponents of υ≥1 and .     

Synchronization in evolving snowdrift game model

The interaction between the evolution of the game and the underlying network structure with evolving snowdrift game model is investigated. The constructed network follows a power-law degree distribution typically showing scale-free feature. The topological features of average path length, clustering coefficient, degree-degree correlations and the dynamical feature of synchronizability are studied. The synchronizability of the constructed networks changes by the interaction. It will converge to a certain value when sufficient new nodes are added. It is found that initial payoffs of nodes greatly affect the synchronizability. When initial payoffs for players are equal, low common initial payoffs may lead to more heterogeneity of the network and good synchronizability. When initial payoffs follow certain distributions, better synchronizability is obtained compared to equal initial payoff. The result is also true for phase synchronization of nonidentical oscillators.     

Assortativeness and information in scale-free networks

We analyze Shannon information of scale-free networks in terms of their assortativeness, and identify classes of networks according to the dependency of the joint remaining degree distribution on the assortativeness. We conjecture that these classes comprise minimalistic and maximalistic networks in terms of Shannon information. For the studied classes, the information is shown to depend non-linearly on the absolute value of the assortativeness, with the dominant term of the relationship being a power-law. We exemplify this dependency using a range of real-world networks. Optimization of scale-free networks according to information they contain depends on the landscape of parameters’ search-space, and we identify two regions of interest: a slope region and a stability region. In the slope region, there is more freedom to generate and evaluate candidate networks since the information content can be changed easily by modifying only the assortativeness, while even a small change in the power-law’s scaling exponent brings a reward in a higher rate of information change. This feature may explain why the exponents of real-world scale-free networks are within a certain range, defined by the slope and stability regions.     

Stylized facts from a threshold-based heterogeneous agent model

A class of heterogeneous agent models is investigated where investors switch trading position whenever their motivation to do so exceeds some critical threshold. These motivations can be psychological in nature or reflect behaviour suggested by the efficient market hypothesis (EMH). By introducing different propensities into a baseline model that displays EMH behaviour, one can attempt to isolate their effects upon the market dynamics. The simulation results indicate that the introduction of a herding propensity results in excess kurtosis and power-law decay consistent with those observed in actual return distributions, but not in significant long-term volatility correlations. Possible alternatives for introducing such long-term volatility correlations are then identified and discussed.     

Local assortativity and growth of Internet

Local assortativity has been recently proposed as a measure to analyse complex networks. It has been noted that the Internet Autonomous System level networks show a markedly different local assortativity profile to most biological and social networks. In this paper we show that, even though several Internet growth models exist, none of them produce the local assortativity profile that can be observed in the real AS networks. We introduce a new generic growth model which can produce a linear local assortativity profile similar to that of the Internet. We verify that this model accurately depicts the local assortativity profile criteria of Internet, while also satisfactorily modelling other attributes of AS networks already explained by existing models.     

Neutron diffraction study of the magnetic structure of Na2 RuO 4

We present a novel model to simulate real social networks of complex interactions, based in a system of colliding particles (agents). The network is build by keeping track of the collisions and evolves in time with correlations which emerge due to the mobility of the agents. Therefore, statistical features are a consequence only of local collisions among its individual agents. Agent dynamics is realized by an event-driven algorithm of collisions where energy is gained as opposed to physical systems which have dissipation. The model reproduces empirical data from networks of sexual interactions, not previously obtained with other approaches.     

Evolution of community structure in the world trade web

In this note we study the bilateral merchandise trade flows between 186 countries over the 1948–2005 period using data from the International Monetary Fund. We use the network visualization package Pajek to identify network structure and behavior across thresholds and over time. In particular, we focus on the evolution of trade “islands” in a world trade network in which countries are linked with directed edges weighted according to the fraction of total dollars sent from one country to another. We find mixed evidence for globalization.     

Detection of node group membership in networks with group overlap

Most networks found in social and biochemical systems have modular structures. An important question prompted by the modularity of these networks is whether nodes can be said to belong to a single group. If they cannot, we would need to consider the role of “overlapping communities.” Despite some efforts in this direction, the problem of detecting overlapping groups remains unsolved because there is neither a formal definition of overlapping community, nor an ensemble of networks with which to test the performance of group detection algorithms when nodes can belong to more than one group. Here, we introduce an ensemble of networks with overlapping groups. We then apply three group identification methods – modularity maximization, k-clique percolation, and modularity-landscape surveying – to these networks. We find that the modularity-landscape surveying method is the only one able to detect heterogeneities in node memberships, and that those heterogeneities are only detectable when the overlap is small. Surprisingly, we find that the k-clique percolation method is unable to detect node membership for the overlapping case.     

Role models for complex networks

We present a framework for automatically decomposing (“block-modeling”) the functional classes of agents within a complex network. These classes are represented by the nodes of an image graph (“block model”) depicting the main patterns of connectivity and thus functional roles in the network. Using a first principles approach, we derive a measure for the fit of a network to any given image graph allowing objective hypothesis testing. From the properties of an optimal fit, we derive how to find the best fitting image graph directly from the network and present a criterion to avoid overfitting. The method can handle both two-mode and one-mode data, directed and undirected as well as weighted networks and allows for different types of links to be dealt with simultaneously. It is non-parametric and computationally efficient. The concepts of structural equivalence and modularity are found as special cases of our approach. We apply our method to the world trade network and analyze the roles individual countries play in the global economy.     

Cascading dynamics in congested complex networks

Cascading failures often occur in congested complex networks. Cascading failures can be expressed as a three phase process: generation, diffusion and dissipation of congestion. Different from betweenness centrality, we propose a congestion function to represent the extent of congestion on a given node. By introducing the concept of “delay time”, we construct an intergradation between permanent removal and nonremoval. We also build a new evaluation function of network efficiency, based on congestion, which measures the damage caused by cascading failures. Finally, based on Statnet and Webgraph topologies we investigate the effects of network structure and size, delay time, processing ability and traffic generation speed on congestion propagation. Also we uncover cascading process composed of three phases and some factors affecting cascade propagation.     

Quantifying bid-ask spreads in the Chinese stock market using limit-order book data

The statistical properties of the bid-ask spread of a frequently traded Chinese stock listed on the Shenzhen Stock Exchange are investigated using the limit-order book data. Three different definitions of spread are considered based on the time right before transactions, the time whenever the highest buying price or the lowest selling price changes, and a fixed time interval. The results are qualitatively similar no matter linear prices or logarithmic prices are used. The average spread exhibits evident intraday patterns consisting of a big L-shape in morning transactions and a small L-shape in the afternoon. The distributions of the spread with different definitions decay as power laws. The tail exponents of spreads at transaction level are well within the interval (2,3) and that of average spreads are well in line with the inverse cubic law for different time intervals. Based on the detrended fluctuation analysis, we found the evidence of long memory in the bid-ask spread time series for all three definitions, even after the removal of the intraday pattern. Using the classical box-counting approach for multifractal analysis, we show that the time series of bid-ask spread do not possess multifractal nature.     

Preferred numbers and the distributions of trade sizes and trading volumes in the Chinese stock market

The distributions of trade sizes and trading volumes are investigated based on the limit order book data of 22 liquid Chinese stocks listed on the Shenzhen Stock Exchange in the whole year 2003. We observe that the size distribution of trades for individualstocks exhibits jumps, which is caused by the number preference of traders when placing orders. We analyze the applicability of the “q-Gamma” function for fitting the distribution by the Cramér-von Mises criterion. The empirical PDFs of tradingvolumes at different timescales Δt ranging from 1 min to 240 min can be well modeled. The applicability of the q-Gamma functions for multiple trades is restricted to the transaction numbers Δn≤ 8. We find that all the PDFs have power-law tails for large volumes. Using careful estimation of the average tail exponents α of the distributions of trade sizes and trading volumes, we get α> 2, well outside the Lévy regime.     

Non-extensive trends in the size distribution of coding and non-coding DNA sequences in the human genome

We study the primary DNA structure of four of the most completely sequenced human chromosomes (including chromosome 19 which is the most dense in coding), using non-extensive statistics. We show that the exponents governing the spatial decay of the coding size distributions vary between 5.2 ≤r ≤5.7 for the short scales and 1.45 ≤q ≤1.50 for the large scales. On the contrary, the exponents governing the spatial decay of the non-coding size distributions in these four chromosomes, take the values 2.4 ≤r ≤3.2 for the short scales and 1.50 ≤q ≤1.72 for the large scales. These results, in particular the values of the tail exponent q, indicate the existence of correlations in the coding and non-coding size distributions with tendency for higher correlations in the non-coding DNA.     

Exact solutions for network rewiring models

Evolving networks with a constant number of edges may be modelled using a rewiring process. These models are used to describe many real-world processes including the evolution of cultural artifacts such as family names, the evolution of gene variations, and the popularity of strategies in simple econophysics models such as the minority game. The model is closely related to Urn models used for glasses, quantum gravity and wealth distributions. The full mean field equation for the degree distribution is found and its exact solution and generating solution are given.     

Elasticity transition and loop formation in vibrated bead chains: A simulation of polymer chains

By measuring the distribution function of the end-to-end distance, we find that strongly shaken bead chains exhibit many properties, such as the rigid-rod-to-Gaussian chain transition, scaling, fast drop of loop formation probability in the short-chain regime, and enhancement of loop formation probability for kinked chains, of long-chain polymers. Though there is difference in local details between our chains and the worm-like chains, our results are consistent with recent calculations based on the worm-like chain model in many respects.