Modelling time-varying interactions in complex systems: the Score Driven Kinetic Ising Model

We introduce a generalization of the Kinetic Ising Model using the score-driven approach, which allows the efficient estimation and filtering of time-varying parameters from time series data. We show that this approach allows to overcome systematic errors in the parameter estimation, and is useful to study complex systems of interacting variables where the strength of the interactions is not constant in time: in particular we propose to quantify the amount of noise in the data and the reliability of forecasts, as well as to discriminate between periods of higher or lower endogeneity in the observed dynamics, namely when interactions are more or less relevant in determining the realization of the observations. We apply our methodology to three different financial settings to showcase some realistic applications, focusing on forecasting high-frequency volatility of stocks, measuring its endogenous component during extreme events in the market, and analysing the strategic behaviour of traders around news releases. We find interesting results on financial systems and, given the widespread use of Ising models in multiple fields, we believe our approach can be efficiently adapted to a variety of settings, ranging from neuroscience to social sciences and machine learning.

[1]  L. Horváth,et al.  Merits and Drawbacks of Variance Targeting in GARCH Models , 2011 .

[2]  Nikolaus Hautsch,et al.  The Impact of Macroeconomic News on Quote Adjustments, Noise, and Informational Volatility , 2010 .

[3]  Fabrizio Lillo,et al.  Inference of the kinetic Ising model with heterogeneous missing data. , 2018, Physical review. E.

[4]  R. W. Penney,et al.  Coupled dynamics of fast spins and slow interactions in neural networks and spin systems , 1993 .

[5]  D. Sornette,et al.  Quantifying Reflexivity in Financial Markets: Towards a Prediction of Flash Crashes , 2012, Physical review. E, Statistical, nonlinear, and soft matter physics.

[6]  M. Tumminello,et al.  How news affects the trading behaviour of different categories of investors in a financial market , 2012, 1207.3300.

[7]  F. Blasques,et al.  Finite Sample Optimality of Score-Driven Volatility Models , 2017 .

[8]  Rémy Chicheportiche,et al.  Trader Lead-Lag Networks and Order Flow Prediction , 2016 .

[9]  MadhavanAnanth,et al.  Exchange-Traded Funds, Market Structure, and the Flash Crash , 2012 .

[10]  P. Anderson More is different. , 1972, Science.

[11]  Didier Sornette,et al.  Apparent criticality and calibration issues in the Hawkes self-excited point process model: application to high-frequency financial data , 2013, 1308.6756.

[12]  S. Koopman,et al.  Predicting Time-Varying Parameters with Parameter-Driven and Observation-Driven Models , 2012 .

[13]  M. Mezard,et al.  Exact mean-field inference in asymmetric kinetic Ising systems , 2011, 1103.3433.

[14]  J. Hertz,et al.  Mean field theory for nonequilibrium network reconstruction. , 2010, Physical review letters.

[15]  T. Bollerslev,et al.  Generalized autoregressive conditional heteroskedasticity , 1986 .

[16]  George Tauchen,et al.  THE PRICE VARIABILITY-VOLUME RELATIONSHIP ON SPECULATIVE MARKETS , 1983 .

[17]  H. Madsen,et al.  Long Memory of Financial Time Series and Hidden Markov Models with Time-Varying Parameters , 2017 .

[18]  Yuguo Chen,et al.  Latent space models for dynamic networks with weighted edges , 2020, Soc. Networks.

[19]  Thierry Mora,et al.  Separating intrinsic interactions from extrinsic correlations in a network of sensory neurons , 2018, Physical Review E.

[20]  S. Bornholdt,et al.  Expectation Bubbles in a Spin Model of Markets , 2001, cond-mat/0105224.

[21]  A. Barra,et al.  A thermodynamic perspective of immune capabilities. , 2011, Journal of theoretical biology.

[22]  Andrew P. Bradley,et al.  The use of the area under the ROC curve in the evaluation of machine learning algorithms , 1997, Pattern Recognit..

[23]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[24]  Eric Budish,et al.  Quantifying the High-Frequency Trading "Arms Race" , 2020, SSRN Electronic Journal.

[25]  Maureen O'Hara,et al.  The Microstructure of the “Flash Crash”: Flow Toxicity, Liquidity Crashes, and the Probability of Informed Trading , 2011, The Journal of Portfolio Management.

[27]  J. Bouchaud Crises and Collective Socio-Economic Phenomena: Simple Models and Challenges , 2012, 1209.0453.

[28]  Alain Destexhe,et al.  Pairwise Ising Model Analysis of Human Cortical Neuron Recordings , 2017, GSI.

[29]  F. Battaglia,et al.  Functional coupling networks inferred from prefrontal cortex activity show experience-related effective plasticity , 2017, Network Neuroscience.

[30]  D. Budescu Dominance analysis: A new approach to the problem of relative importance of predictors in multiple regression. , 1993 .

[31]  Jean-Philippe Bouchaud,et al.  Branching-ratio approximation for the self-exciting Hawkes process. , 2014, Physical review. E, Statistical, nonlinear, and soft matter physics.

[32]  Marco P. Tucci,et al.  Time-varying parameters: a critical introduction , 1995 .

[33]  J. Hanley,et al.  The meaning and use of the area under a receiver operating characteristic (ROC) curve. , 1982, Radiology.

[34]  E. Gardner,et al.  An Exactly Solvable Asymmetric Neural Network Model , 1987 .

[35]  O. Schmitz Predator and prey functional traits: understanding the adaptive machinery driving predator–prey interactions , 2017, F1000Research.

[36]  H. White,et al.  Advances in Econometrics: Specification testing in dynamic models , 1987 .

[37]  H. Manner,et al.  Dynamic stochastic copula models: Estimation, inference and applications , 2012 .

[38]  Tomasz Górski,et al.  Cholinergic switch between two different types of slow waves in cerebral cortex , 2018, bioRxiv.

[39]  D. Sornette Physics and financial economics (1776–2014): puzzles, Ising and agent-based models , 2014, Reports on progress in physics. Physical Society.

[40]  A. Harvey Dynamic Models for Volatility and Heavy Tails: With Applications to Financial and Economic Time Series , 2013 .

[41]  Fabrizio Lillo,et al.  A dynamic network model with persistent links and node-specific latent variables, with an application to the interbank market , 2017, Eur. J. Oper. Res..

[42]  L. Bauwens,et al.  The stochastic conditional duration model: a latent variable model for the analysis of financial durations , 2004 .

[43]  A. Kyle,et al.  The Flash Crash: High-Frequency Trading in an Electronic Market , 2017 .

[44]  Siem Jan Koopman,et al.  Information-theoretic optimality of observation-driven time series models for continuous responses , 2015 .

[45]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[46]  Y. Bar-Yam General Features of Complex Systems , 2011 .

[47]  F. Lillo,et al.  Detection of intensity bursts using Hawkes processes: An application to high-frequency financial data. , 2016, Physical Review E.

[48]  A. Pelizzola,et al.  Ising-like model for protein mechanical unfolding. , 2006, Physical review letters.

[49]  D. Budescu,et al.  The dominance analysis approach for comparing predictors in multiple regression. , 2003, Psychological methods.

[50]  Marc Barthelemy,et al.  The multilayer temporal network of public transport in Great Britain , 2015, Scientific Data.

[51]  Jonathan W. Schooler,et al.  States of Mind: Characterizing the Neural Bases of Focus and Mind-wandering through Dynamic Functional Connectivity , 2017, Journal of Cognitive Neuroscience.

[52]  Stephen J. Hardiman,et al.  Critical reflexivity in financial markets: a Hawkes process analysis , 2013, 1302.1405.

[53]  H. Scheraga,et al.  Model of protein folding: incorporation of a one-dimensional short-range (Ising) model into a three-dimensional model. , 1977, Proceedings of the National Academy of Sciences of the United States of America.

[54]  Christian Beck,et al.  From time series to superstatistics. , 2005, Physical review. E, Statistical, nonlinear, and soft matter physics.

[55]  D. Sornette,et al.  Quantifying reflexivity in financial markets: towards a prediction of flash crashes , 2012, 1201.3572.

[56]  Pan Zhang,et al.  Inference of the sparse kinetic Ising model using the decimation method. , 2015, Physical review. E, Statistical, nonlinear, and soft matter physics.

[57]  O. Marre,et al.  Maximum-entropy models reveal the excitatory and inhibitory correlation structures in cortical neuronal activity. , 2018, Physical review. E.

[58]  Zoubin Ghahramani,et al.  An Introduction to Hidden Markov Models and Bayesian Networks , 2001, Int. J. Pattern Recognit. Artif. Intell..

[59]  Giuseppe Cavaliere Stochastic Volatility: Selected Readings , 2006 .

[60]  Sompolinsky,et al.  Dynamics of spin systems with randomly asymmetric bonds: Ising spins and Glauber dynamics. , 1988, Physical review. A, General physics.

[61]  D. Sornette,et al.  The Endo-Exo Problem in High Frequency Financial Price Fluctuations and Rejecting Criticality , 2018 .

[62]  Drew D. Creal,et al.  Testing for Parameter Instability across Different Modeling Frameworks , 2016 .

[63]  Albert J. Menkveld,et al.  The Flash Crash: A Cautionary Tale about Highly Fragmented Markets , 2017, Manag. Sci..

[64]  Simona Cocco,et al.  Functional networks from inverse modeling of neural population activity , 2017 .

[65]  Fabrizio Lillo,et al.  Modeling foreign exchange market activity around macroeconomic news: Hawkes-process approach. , 2014, Physical review. E, Statistical, nonlinear, and soft matter physics.

[66]  Jari Saramäki,et al.  Temporal Networks , 2011, Encyclopedia of Social Network Analysis and Mining.

[67]  Drew D. Creal,et al.  Generalized autoregressive score models with applications ∗ , 2010 .