Correspondence between temporal correlations in time series, inverse problems, and the Spherical Model

Abstract

In this paper we employ methods from Statistical Mechanics to model temporal correlations in time series. We put forward a methodology based on the Maximum Entropy principle to generate ensembles of time series constrained to preserve part of the temporal structure of an empirical time series of interest. We show that a constraint on the lag-one autocorrelation can be fully handled analytically, and corresponds to the well known Spherical Model of a ferromagnet. We then extend such a model to include constraints on more complex temporal correlations by means of perturbation theory, showing that this leads to substantial improvements in capturing the lag-one autocorrelation in the variance. We apply our approach on synthetic data, and illustrate how it can be used to formulate expectations on the future values of a data generating process.

arXiv comment9 pages, 2 figures

References in corpus (16)

• Statistical physics of social dynamics
• The Statistical Physics of Real-World Networks
• Inverse statistical problems: from the inverse Ising problem to data science
• Statistical mechanics of systems with heterogeneous agents: Minority Games
• The Ising Model for Neural Data: Model Quality and Approximate Methods for Extracting Functional Connectivity
• Maximum likelihood: extracting unbiased information from complex networks
• Dynamical Hurst exponent as a tool to monitor unstable periods in financial time series
• Evolutionary game theory in growing populations
• Maximum Caliber: a general variational principle for dynamical systems
• Interdisciplinarity: A Nobel Opportunity
• A parametric approach to information filtering in complex networks: The Pólya filter
• Taking census of physics
• The evolution of interdisciplinarity in physics research
• Markov processes follow from the principle of Maximum Caliber
• Statistical mechanics of complex economies
• Maximum Entropy approach to multivariate time series randomization

Cited by in corpus (3)

• On the equivalence between the Kinetic Ising Model and discrete autoregressive processes
• Description length of canonical and microcanonical models
• Stochastic gradient descent with random learning rate