English

Finite-size scaling analysis of binary stochastic processes and universality classes of information cascade phase transition

Statistical Mechanics 2015-06-12 v3

Abstract

We propose a finite-size scaling analysis of binary stochastic processes X(t){0,1}X(t)\in \{0,1\} based on the second moment correlation length ξ\xi for the autocorrelation function C(t)C(t). The purpose is to clarify the critical properties and provide a new data analysis method for information cascades. As a simple model to represent the different behaviors of subjects in information cascade experiments, we assume that X(t)X(t) is a mixture of an independent random variable that takes 1 with probability qq and a random variable that depends on the ratio zz of the variables taking 1 among recent rr variables. We consider two types of the probability f(z)f(z) that the latter takes 1: (i) analog [f(z)=zf(z)=z] and (ii) digital [f(z)=θ(z1/2)f(z)=\theta(z-1/2)]. We study the universal functions of scaling for ξ\xi and the integrated correlation time τ\tau. For finite rr, C(t)C(t) decays exponentially as a function of tt, and there is only one stable renormalization group (RG) fixed point. In the limit rr\to \infty, where X(t)X(t) depends on all the previous variables, C(t)C(t) in model (i) obeys a power law, and the system becomes scale invariant. In model (ii) with q1/2q\neq 1/2, there are two stable RG fixed points, which correspond to the ordered and disordered phases of the information cascade phase transition with critical exponents β=1\beta=1 and ν=2\nu_{||}=2.

Keywords

Cite

@article{arxiv.1404.4921,
  title  = {Finite-size scaling analysis of binary stochastic processes and universality classes of information cascade phase transition},
  author = {Shintaro Mori and Masato Hisakado},
  journal= {arXiv preprint arXiv:1404.4921},
  year   = {2015}
}

Comments

32 pages, 9 figures