A Gaussian Mixture Model Based Cost Function for Parameter Estimation of CHaotic Biological Systems

Yasser Shekofteh^a,b, Sajad Jafari^a, Julien Clinton Sprott^c, S. Mohammad Reza Hashemi Golpayegani^a, Farshad Almasganj^a
^aBiomedical Engineering Department, Amirkabir University of Technology, Tehran 15875-4413, Iran
^bResearch Center of Intelligent Signal Processing (RCISP), Tehran, Iran
^cPhysics Department, University of Wisconsin, 1150 University Ave., Madison, WI 53706, USA

Received 21 November 2012, Received in revised form 21 November 2013, Accepted 24 May 2014, Available online 6 June 2014

ABSTRACT

As we know, many biological systems such as neurons or the heart can exhibit chaotic behavior. Conventional methods for parameter estimation in models of these systems have some limitations caused by sensitivity to initial conditions. In this paper, a novel cost function is proposed to overcome those limitations by building a statistical model on the distribution of the real system attractor in state space. This cost function is defined by the use of a likelihood score in a Gaussian mixture model (GMM) which is fitted to the observed attractor generated by the real system. Using that learned GMM, a similarity score can be defined by the computed likelihood score of the model time series. We have applied the proposed method to the parameter estimation of two important biological systems, a neuron and a cardiac pacemaker, which show chaotic behavior. Some simulated experiments are given to verify the usefulness of the proposed approach in clean and noisy conditions. The results show the adequacy of the proposed cost function.

Ref: Y. Shekofteh, S. Jafari, J. C. Sprott, S. M. R. H. Golpayegani, and F. Almasganj, Commun. Nonlinear Sci. Numer. Simulat. 20, 469-481 (2015)

The complete paper is available in PDF format.

Return to Sprott's Books and Publications.