arXiv subjects

Kimmo Kaski

Publications and source records attributed to Kimmo Kaski.

At least 109 records · Page 6Linked to original sources

arXiv2010

Reciprocity of mobile phone calls

We present a study of the reciprocity of human behaviour based on mobile phone usage records. The underlying question is whether human relationships are mutual, in the sense that both are equally active in keeping up the relationship, or is it on the contrary typical that relationships are lopsided, with one party being significantly more active than the other. We study this question with the help of a mobile phone data set consisting of all mobile phone calls between 5.3 million customers of a single mobile phone operator. It turns out that lopsided relations are indeed quite common, to the extent that the variation cannot be explained by simple random deviations or by variations in personal activity. We also show that there is no non-trivial correlation between reciprocity and local network density.

physics.soc-ph↗

arXiv2008

A comparative study of social network models: network evolution models and nodal attribute models

This paper reviews, classifies and compares recent models for social networks that have mainly been published within the physics-oriented complex networks literature. The models fall into two categories: those in which the addition of new links is dependent on the (typically local) network structure (network evolution models, NEMs), and those in which links are generated based only on nodal attributes (nodal attribute models, NAMs). An exponential random graph model (ERGM) with structural dependencies is included for comparison. We fit models from each of these categories to two empirical acquaintance networks with respect to basic network properties. We compare higher order structures in the resulting networks with those in the data, with the aim of determining which models produce the most realistic network structure with respect to degree distributions, assortativity, clustering spectra, geodesic path distributions, and community structure (subgroups with dense internal connections). We find that the nodal attribute models successfully produce assortative networks and very clear community structure. However, they generate unrealistic clustering spectra and peaked degree distributions that do not match empirical data on large social networks. On the other hand, many of the network evolution models produce degree distributions and clustering spectra that agree more closely with data. They also generate assortative networks and community structure, although often not to the same extent as in the data. The ERG model turns out to produce the weakest community structure.

physics.soc-ph↗

arXiv2008

Maximal spanning trees, asset graphs and random matrix denoising in the analysis of dynamics of financial networks

We study the time dependence of maximal spanning trees and asset graphs based on correlation matrices of stock returns. In these networks the nodes represent companies and links are related to the correlation coefficients between them. Special emphasis is given to the comparison between ordinary and denoised correlation matrices. The analysis of single- and multi-step survival ratios of the corresponding networks reveals that the ordinary correlation matrices are more stable in time than the denoised ones. Our study also shows that some information about the cluster structure of the companies is lost in the denoising procedure. Cluster structure that makes sense from an economic point of view exists, and can easily be observed in networks based on denoised correlation matrices. However, this structure is somewhat clearer in the networks based on ordinary correlation matrices. Some technical aspects, such as the random matrix denoising procedure, are also presented.

physics.soc-ph↗

arXiv2008

Detecting modules in dense weighted networks with the Potts method

We address the problem of multiresolution module detection in dense weighted networks, where the modular structure is encoded in the weights rather than topology. We discuss a weighted version of the q-state Potts method, which was originally introduced by Reichardt and Bornholdt. This weighted method can be directly applied to dense networks. We discuss the dependence of the resolution of the method on its tuning parameter and network properties, using sparse and dense weighted networks with built-in modules as example cases. Finally, we apply the method to data on stock price correlations, and show that the resulting modules correspond well to known structural properties of this correlation network.

physics.soc-ph↗

arXiv2008

A sequential algorithm for fast clique percolation

In complex network research clique percolation, introduced by Palla et al., is a deterministic community detection method, which allows for overlapping communities and is purely based on local topological properties of a network. Here we present a sequential clique percolation algorithm (SCP) to do fast community detection in weighted and unweighted networks, for cliques of a chosen size. This method is based on sequentially inserting the constituent links to the network and simultaneously keeping track of the emerging community structure. Unlike existing algorithms, the SCP method allows for detecting k-clique communities at multiple weight thresholds in a single run, and can simultaneously produce a dendrogram representation of hierarchical community structure. In sparse weighted networks, the SCP algorithm can also be used for implementing the weighted clique percolation method recently introduced by Farkas et al. The computational time of the SCP algorithm scales linearly with the number of k-cliques in the network. As an example, the method is applied to a product association network, revealing its nested community structure.

physics.soc-ph↗

arXiv2007

Spectral methods and cluster structure in correlation-based networks

We investigate how in complex systems the eigenpairs of the matrices derived from the correlations of multichannel observations reflect the cluster structure of the underlying networks. For this we use daily return data from the NYSE and focus specifically on the spectral properties of weight W_{ij} = |C|_{ij} - δ_{ij} and diffusion matrices D_{ij} = W_{ij}/s_j- δ_{ij}, where C_{ij} is the correlation matrix and s_i = \sum_j W_{ij} the strength of node j. The eigenvalues (and corresponding eigenvectors) of the weight matrix are ranked in descending order. In accord with the earlier observations the first eigenvector stands for a measure of the market correlations. Its components are to first approximation equal to the strengths of the nodes and there is a second order, roughly linear, correction. The high ranking eigenvectors, excluding the highest ranking one, are usually assigned to market sectors and industrial branches. Our study shows that both for weight and diffusion matrices the eigenpair analysis is not capable of easily deducing the cluster structure of the network without a priori knowledge. In addition we have studied the clustering of stocks using the asset graph approach with and without spectrum based noise filtering. It turns out that asset graphs are quite insensitive to noise and there is no sharp percolation transition as a function of the ratio of bonds included, thus no natural threshold value for that ratio seems to exist. We suggest that these observations can be of use for other correlation based networks as well.

physics.soc-ph↗

arXiv2007

Limited resolution in complex network community detection with Potts model approach

According to Fortunato and Barthelemy, modularity-based community detection algorithms have a resolution threshold such that small communities in a large network are invisible. Here we generalize their work and show that the q-state Potts community detection method introduced by Reichardt and Bornholdt also has a resolution threshold. The model contains a parameter by which this threshold can be tuned, but no a priori principle is known to select the proper value. Single global optimization criteria do not seem capable for detecting all communities if their size distribution is broad.

cond-mat.dis-nn↗

arXiv2007

Spectral and network methods in the analysis of correlation matrices of stock returns

Correlation matrices inferred from stock return time series contain information on the behaviour of the market, especially on clusters of highly correlating stocks. Here we study a subset of New York Stock Exchange (NYSE) traded stocks and compare three different methods of analysis: i) spectral analysis, i.e. investigation of the eigenvalue-eigenvector pairs of the correlation matrix, ii) asset trees, obtained by constructing the maximal spanning tree of the correlation matrix, and iii) asset graphs, which are networks in which the strongest correlations are depicted as edges. We illustrate and discuss the localisation of the most significant modes of fluctuation, i.e. eigenvectors corresponding to the largest eigenvalues, on the asset trees and graphs.

physics.soc-ph↗

arXiv2007

Analysis of a large-scale weighted network of one-to-one human communication

We construct a connected network of 3.9 million nodes from mobile phone call records, which can be regarded as a proxy for the underlying human communication network at the societal level. We assign two weights on each edge to reflect the strength of social interaction, which are the aggregate call duration and the cumulative number of calls placed between the individuals over a period of 18 weeks. We present a detailed analysis of this weighted network by examining its degree, strength, and weight distributions, as well as its topological assortativity and weighted assortativity, clustering and weighted clustering, together with correlations between these quantities. We give an account of motif intensity and coherence distributions and compare them to a randomized reference system. We also use the concept of link overlap to measure the number of common neighbors any two adjacent nodes have, which serves as a useful local measure for identifying the interconnectedness of communities. We report a positive correlation between the overlap and weight of a link, thus providing strong quantitative evidence for the weak ties hypothesis, a central concept in social network analysis. The percolation properties of the network are found to depend on the type and order of removed links, and they can help understand how the local structure of the network manifests itself at the global level. We hope that our results will contribute to modeling weighted large-scale social networks, and believe that the systematic approach followed here can be adopted to study other weighted networks.

physics.soc-ph↗

arXiv2007

Emergence of self-sustained patterns in small-world excitable media

Motivated by recent observations that long-range connections (LRC) play a role in various brain phenomena, we have observed two distinct dynamical transitions in the activity of excitable media where waves propagate both between neighboring regions and through LRC. When LRC density $p$ is low, single or multiple spiral waves are seen to emerge and cover the entire system. This state is self-sustaining and robust against perturbations. At $p=p_c^l$ the spirals are suppressed and there is a transition to a spatially homogeneous, temporally periodic state. Finally, above $p=p_c^u$, activity ceases after a brief transient.

cond-mat.dis-nn↗

arXiv2005

Kinetic theory models for the distribution of wealth: power law from overlap of exponentials

Various multi-agent models of wealth distributions defined by microscopic laws regulating the trades, with or without a saving criterion, are reviewed. We discuss and clarify the equilibrium properties of the model with constant global saving propensity, resulting in Gamma distributions, and their equivalence to the Maxwell-Boltzmann kinetic energy distribution for a system of molecules in an effective number of dimensions $D_λ$, related to the saving propensity $λ$ [M. Patriarca, A. Chakraborti, and K. Kaski, Phys. Rev. E 70 (2004) 016104]. We use these results to analyze the model in which the individual saving propensities of the agents are quenched random variables, and the tail of the equilibrium wealth distribution exhibits a Pareto law $f(x) \propto x^{-α-1}$ with an exponent $α=1$ [A. Chatterjee, B. K. Chakrabarti, and S. S. Manna, Physica Scripta T106 (2003) 367]. Here, we show that the observed Pareto power law can be explained as arising from the overlap of the Maxwell-Boltzmann distributions associated to the various agents, which reach an equilibrium state characterized by their individual Gamma distributions. We also consider the influence of different types of saving propensity distributions on the equilibrium state.

physics.soc-ph↗

arXiv2005

Spatial snowdrift game with myopic agents

We have studied a spatially extended snowdrift game, in which the players are located on the sites of two-dimensional square lattices and repeatedly have to choose one of the two strategies, either cooperation (C) or defection (D). A player interacts with its nearest neighbors only, and aims at playing a strategy which maximizes its instant pay-off, assuming that the neighboring agents retain their strategies. If a player is not content with its current strategy, it will change it to the opposite one with probability $p$ next round. Here we show through simulations and analytical approach that these rules result in cooperation levels, which differ to large extent from those obtained using the replicator dynamics.

physics.bio-ph↗

arXiv2004

Invisible Hand Effect in an Evolutionary Minority Game Model

In this paper we study the properties of a Minority Game with evolution realized by using genetic crossover to modify fixed-length decision-making strategies of agents. Although the agents in this evolutionary game act selfishly by trying to maximize their own performances only, it turns out that the whole society will eventually be rewarded optimally. This "Invisible Hand" effect is what Adam Smith over two centuries ago expected to take place in the context of free market mechanism. However, this behaviour of the society of agents is realized only under idealized conditions, where all agents are utilizing the same efficient evolutionary mechanism. If on the other hand part of the agents are adaptive, but not evolutionary, the system does not reach optimum performance, which is also the case if part of the evolutionary agents form a uniformly acting "cartel".

cond-mat.other↗

arXiv2004

Scale-Free Networks Generated By Random Walkers

We present a simple mechanism for generating undirected scale-free networks using random walkers, where the network growth is determined by choosing parent vertices by sequential random walks. We show that this mechanism produces scale-free networks with degree exponent gamma=3 and clustering coefficients depending on random walk length. The mechanism can be interpreted in terms of preferential attachment without explicit knowledge of node degrees.

cond-mat.stat-mech↗

arXiv2004

A statistical model with a standard Gamma distribution

We study a statistical model consisting of $N$ basic units which interact with each other by exchanging a physical entity, according to a given microscopic random law, depending on a parameter $λ$. We focus on the equilibrium or stationary distribution of the entity exchanged and verify through numerical fitting of the simulation data that the final form of the equilibrium distribution is that of a standard Gamma distribution. The model can be interpreted as a simple closed economy in which economic agents trade money and a saving criterion is fixed by the saving propensity $λ$. Alternatively, from the nature of the equilibrium distribution, we show that the model can also be interpreted as a perfect gas at an effective temperature $T(λ)$, where particles exchange energy in a space with an effective dimension $D(λ)$.

cond-mat.stat-mech↗

arXiv2003

Gibbs versus non-Gibbs distributions in money dynamics

We review a simple model of closed economy, where the economic agents make money transactions and a saving criterion is present. We observe the Gibbs distribution for zero saving propensity, and non-Gibbs distributions otherwise. While the exact solution in the case of zero saving propensity is already known to be given by the Gibbs distribution, here we provide the explicit analytical form of the equilibrium distribution for the general case of nonzero saving propensity. We verify it through comparison with numerical data and show that it can be cast in the form of a Poisson distribution.

cond-mat.stat-mech↗

arXiv2003

Dimensionality effects in Turing pattern formation

The problem of morphogenesis and Turing instability are revisited from the point of view of dimensionality effects. First the linear analysis of a generic Turing model is elaborated to the case of multiple stationary states, which may lead the system to bistability. The difference between two- and three-dimensional pattern formation with respect to pattern selection and robustness is discussed. Preliminary results concerning the transition between quasi-two-dimensional and three-dimensional structures are presented and their relation to experimental results are addressed.

cond-mat.soft↗

arXiv2003

Searching good strategies in adaptive minority games

In this paper we introduce adaptation mechanism based on genetic algorithms in minority games. If agents find their performances too low, they modify their strategies in hope to improve their performances and become more successful. One aim of this study is to find out what happens at the system as well as at the individual agent level. We observe that adaptation remarkably tightens the competition among the agents, and tries to pull the collective system into a state where the aggregate utility is the largest. We first make a brief comparative study of the different adaptation mechanisms and then present in more detail parametric studies. These different adaptation mechanisms broaden the scope of the applications of minority games to the study of complex systems.

cond-mat.stat-mech↗