arXiv subjects

Panos Argyrakis

Publications and source records attributed to Panos Argyrakis.

31 records · Page 2Linked to original sources

arXiv2008

The structural role of weak and strong links in a financial market network

We investigate the properties of correlation based networks originating from economic complex systems, such as the network of stocks traded at the New York Stock Exchange (NYSE). The weaker links (low correlation) of the system are found to contribute to the overall connectivity of the network significantly more than the strong links (high correlation). We find that nodes connected through strong links form well defined communities. These communities are clustered together in more complex ways compared to the widely used classification according to the economic activity. We find that some companies, such as General Electric (GE), Coca Cola (KO), and others, can be involved in different communities. The communities are found to be quite stable over time. Similar results were obtained by investigating markets completely different in size and properties, such as the Athens Stock Exchange (ASE). The present method may be also useful for other networks generated through correlations.

q-fin.ST↗

arXiv2007

Scale-free networks resistant to intentional attacks

We study the detailed mechanism of the failure of scale-free networks under intentional attacks. Although it is generally accepted that such networks are very sensitive to targeted attacks, we show that for a particular type of structure such networks surprisingly remain very robust even under removal of a large fraction of their nodes, which in some cases can be up to 70%. The degree distribution $P(k)$ of these structures is such that for small values of the degree $k$ the distribution is constant with $k$, up to a critical value $k_c$, and thereafter it decays with $k$ with the usual power law. We describe in detail a model for such a scale-free network with this modified degree distribution, and we show both analytically and via simulations, that this model can adequately describe all the features and breakdown characteristics of these attacks. We have found several experimental networks with such features, such as for example the IMDB actors collaboration network or the citations network, whose resilience to attacks can be accurately described by our model.

physics.soc-ph↗

arXiv2007

Priority diffusion model in lattices and complex networks

We introduce a model for diffusion of two classes of particles ($A$ and $B$) with priority: where both species are present in the same site the motion of $A$'s takes precedence over that of $B$'s. This describes realistic situations in wireless and communication networks. In regular lattices the diffusion of the two species is normal but the $B$ particles are significantly slower, due to the presence of the $A$ particles. From the fraction of sites where the $B$ particles can move freely, which we compute analytically, we derive the diffusion coefficients of the two species. In heterogeneous networks the fraction of sites where $B$ is free decreases exponentially with the degree of the sites. This, coupled with accumulation of particles in high-degree nodes leads to trapping of the low priority particles in scale-free networks.

cond-mat.stat-mech↗

arXiv2007

Improving immunization strategies

We introduce an immunization method where the percentage of required vaccinations for immunity are close to the optimal value of a targeted immunization scheme of highest degree nodes. Our strategy retains the advantage of being purely local, without the need of knowledge on the global network structure or identification of the highest degree nodes. The method consists of selecting a random node and asking for a neighbor that has more links than himself or more than a given threshold and immunizing him. We compare this method to other efficient strategies on three real social networks and on a scale-free network model, and find it to be significantly more effective.

physics.soc-ph↗

arXiv2007

Evolution of Vocabulary on Scale-free and Random Networks

We examine the evolution of the vocabulary of a group of individuals (linguistic agents) on a scale-free network, using Monte Carlo simulations and assumptions from evolutionary game theory. It is known that when the agents are arranged in a two-dimensional lattice structure and interact by diffusion and encounter, then their final vocabulary size is the maximum possible. Knowing all available words is essential in order to increase the probability to ``survive'' by effective reproduction. On scale-free networks we find a different result. It is not necessary to learn the entire vocabulary available. Survival chances are increased by using the vocabulary of the ``hubs'' (nodes with high degree). The existence of the ``hubs'' in a scale-free network is the source of an additional important fitness generating mechanism.

physics.soc-ph↗

arXiv2006

Characteristics of reaction-diffusion on scale-free networks

We examine some characteristic properties of reaction-diffusion processes of the A+A->0 type on scale-free networks. Due to the inhomogeneity of the structure of the substrate, as compared to usual lattices, we focus on the characteristics of the nodes where the annihilations occur. We show that at early times the majority of these events take place on low-connectivity nodes, while as time advances the process moves towards the high-connectivity nodes, the so-called hubs. This pattern remarkably accelerates the annihilation of the particles, and it is in agreement with earlier predictions that the rates of reaction-diffusion processes on scale-free networks are much faster than the equivalent ones on lattice systems.

cond-mat.dis-nn↗

arXiv2006

Language Time Series Analysis

We use the Detrended Fluctuation Analysis (DFA) and the Grassberger-Proccacia analysis (GP) methods in order to study language characteristics. Despite that we construct our signals using only word lengths or word frequencies, excluding in this way huge amount of information from language, the application of Grassberger- Proccacia (GP) analysis indicates that linguistic signals may be considered as the manifestation of a complex system of high dimensionality, different from random signals or systems of low dimensionality such as the earth climate. The DFA method is additionally able to distinguish a natural language signal from a computer code signal. This last result may be useful in the field of cryptography.

physics.soc-ph↗

arXiv2005

Statistical Mechanical Approach to Human Language

We use the formulation of equilibrium statistical mechanics in order to study some important characteristics of language. Using a simple expression for the Hamiltonian of a language system, which is directly implied by the Zipf law, we are able to explain several characteristic features of human language that seem completely unrelated, such as the universality of the Zipf exponent, the vocabulary size of children, the reduced communication abilities of people suffering from schizophrenia, etc. While several explanations are necessarily only qualitative at this stage, we have, nevertheless, been able to derive a formula for the vocabulary size of children as a function of age, which agrees rather well with experimental data.

physics.soc-ph↗

arXiv2005

Stability and topology of scale-free networks under attack and defense strategies

We study tolerance and topology of random scale-free networks under attack and defense strategies that depend on the degree k of the nodes. This situation occurs, for example, when the robustness of a node depends on its degree or in an intentional attack with insufficient knowledge on the network. We determine, for all strategies, the critical fraction p_c of nodes that must be removed for disintegrating the network. We find that for an intentional attack, little knowledge of the well-connected sites is sufficient to strongly reduce p_c. At criticality, the topology of the network depends on the removal strategy, implying that different strategies may lead to different kinds of percolation transitions.

cond-mat.dis-nn↗

arXiv2005

Reaction-diffusion processes on correlated and uncorrelated scale-free networks

We compare reaction-diffusion processes of the $A+A\to 0$ type on scale-free networks created with either the configuration model or the uncorrelated configuration model. We show via simulations that except for the difference in the behavior of the two models, different results are observed within the same model when the minimum number of connections for a node varies from $k_{\rm min}=1$ to $k_{\rm min}=2$. This difference is attributed to the varying local properties of the two systems. In all cases we are able to identify a power law behavior of the density decay with time with an exponent $f>1$, considerably larger than its lattice counterpart.

cond-mat.dis-nn↗

arXiv2005

Language evolution and population Dynamics in a system of two interacting species

We use Monte Carlo simulations and assumptions from evolutionary game theory in order to study the evolution of words and the population dynamics of a system comprising two interacting species which initially speak two different languages. The species are characterized by their identity, vocabulary and have different initial fitness, i.e. reproduction capability. The questions we want to answer are: a. Will the different initial fitness lead to a permanent advantage? b. Will this advantage affect the vocabulary of the species or the population dynamics? c. How will the spatial distributions of the species be affected? Does the system exhibit pattern formation or segregation? We show that an initial fitness advantage, although is very quickly balanced, leads to better spatial arrangement and enhances survival probabilities of the species. In most cases the system will arrive at a final state where both languages coexist. However, in cases where one species greatly outnumbers the other in population and fitness, then only one species survives with its final language having a slightly richer vocabulary than its initial language. Thus, our results offer an explanation for the existence and origin of synonyms in all currently spoken languages.

cond-mat.stat-mech↗

arXiv2004

Absence of kinetic effects in reaction-diffusion processes in scale-free networks

We show that the chemical reactions of the model systems of A+A->0 and A+B->0 when performed on scale-free networks exhibit drastically different behavior as compared to the same reactions in normal spaces. The exponents characterizing the density evolution as a function of time are considerably higher than 1, implying that both reactions occur at a much faster rate. This is due to the fact that the discerning effects of the generation of a depletion zone (A+A) and the segregation of the reactants (A+B) do not occur at all as in normal spaces. Instead we observe the formation of clusters of A (A+A reaction) and of mixed A and B (A+B reaction) around the hubs of the network. Only at the limit of very sparse networks is the usual behavior recovered.

cond-mat.dis-nn↗

arXiv2003

Distribution of infected mass in disease spreading in scale-free networks

We use scale-free networks to study properties of the infected mass $M$ of the network during a spreading process as a function of the infection probability $q$ and the structural scaling exponent $γ$. We use the standard SIR model and investigate in detail the distribution of $M$, We find that for dense networks this function is bimodal, while for sparse networks it is a smoothly decreasing function, with the distinction between the two being a function of $q$. We thus recover the full crossover transition from one case to the other. This has a result that on the same network a disease may die out immediately or persist for a considerable time, depending on the initial point where it was originated. Thus, we show that the disease evolution is significantly influenced by the structure of the underlying population.

cond-mat.dis-nn↗