Spectral methods and cluster structure in correlation-based networks

Abstract

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} - \delta_{ij} and diffusion matrices D_{ij} = W_{ij}/s_j- \delta_{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.Comment: 26 pages, 14 figure

Community Landscapes: An Integrative Approach to Determine Overlapping Network Module Hierarchy, Identify Key Nodes and Predict Network Dynamics

“…N.A. Barnes, 1982;Donath and Hoffman, 1972;Kleinberg, 1997;Gibson et al, 1998;Hartuv and Shamir, 2000;Tibély et al, 2006;Heimo et al, 2008a;Sahai et al, 2009 Division optimization based on community sub-matrix eigenvalue maximization Chauhan et al, 2009 Truncated singular value decomposition of the modular contribution matrix Arenas et al, 2010 Zhang et al, 2007c;Wang et al, 2008b;Ma et al, 2010 Spectral properties of the complement graph…”

Correlations, risk and crisis: From physiology to finance

“…Recently [47] eigenvalues and eigenvectors of the matrix of the absolute values of the correlation coefficients were used for analysis of the New York Stock Exchange (NYSE) traded stocks. The transformation from the correlation matrix to the matrix of absolute values was justified by interpreting the absolute values as measures of interaction strength without considering whether the interaction is positive or negative.…”

Correlations, Risk and Crisis: From Physiology to Finance

“…Liebig's systems of factors violate the synergy inequality (22): if at points f, g with the same values of fitness W (f) = W (g) different factors are limiting, then at the average point the value of both these factors are smaller, and the harm of the limiting factor at that point is less, than at both points f, g, i.e. the fitness at the average point is larger.…”

“…This invariance with respect to nonlinear change of scale is very important, because usually we don't know the values of function W .Proposition 2. If the synergy inequality(22) holds for a function W , then it holds for a function W θ = θ(W ), where θ(x) is an arbitrary strictly monotonic function of one variable. this property allows us to study the problem about optimal distribution of the adaptation resource without further knowledge about the fitness function.Assume that adaptation should maximize an objective function W (f 1 − r 1 , ...f q − r q ) (16) which satisfies the synergy inequality(22) under conditions r…”