基于经验动态建模的因果检测研究综述

doi:10.11896/jsjkx.220600194

摘要/Abstract

摘要： 相关性是目前科学研究中重要的分析标准,但是相关关系并不意味着因果关系。随着人们认识到非线性动力学的普遍性,仅仅依靠相关性进行推论,极有可能会得出错误的结论。目前,包括机器学习在内的各种相关性研究算法快速发展,而挖掘变量间因果关联的研究尚处于探索中。经验动态建模理论是一种基于数据驱动的动态系统建模框架,其最大的特点是抛弃了传统数据分析中的公式化方法,仅仅从时间序列中重构动态系统的行为。其核心思想是动态系统可以描述为一组状态在高维空间中受一定规则的驱使随时间演变的过程,可以通过重建随时间演变的状态来对动态系统进行建模。基于经验动态建模理论,通过动态系统中变量的时序数据,可以对变量之间的因果关系进行检测。如果变量X是变量Y的原因(X⇒Y),那么变量X的信息必须隐含在变量Y中,且可以从变量Y中恢复。文中首先分析了相关性与因果关系的辩证关系,由于“虚假相关性”和“幻象性相关”现象的存在,相关性并不意味着因果性,缺乏相关性也并不意味着没有因果关联;然后全面介绍基于经验动态建模理论进行因果检测的核心思想,具体地对Takens嵌入定理、单纯形投影算法和收敛交叉映射算法的历史发展做了详细总结,动态系统可以描述为一组状态在高维空间中受一定规则的驱使随时间演变的过程,可以通过重建随时间演变的状态来对动态系统进行建模;接着介绍了一些经验动态建模理论改进方法和实际因果检测应用;最后分析了基于经验动态建模因果检测目前存在的一些问题,并对未来发展趋势进行了展望。

关键词: 因果检测, 经验动态建模, 收敛交叉映射, 复杂系统, 人工智能

Abstract: Correlation is an important analysis standard in curent scientific research,but it does not mean causality.As people rea-lize the universality of no nlinear dynamics,it is very likely to lead to wrong conclusions only by relying on corelation.At present,various correlation research algorithms,including machine learning,are developing rapidly,while the research of mining causal correlation between variables is still under exploration.Empirical dynamic modeling theory is a data-driven dynamic system mo-deling framework.Its biggest feature is to abandon the formulaic method in traditional data analysis and reconstruct the behavior of dynamic system only from time series.The core idea is that a dynamic system can be described as a process in which a group of states evolve over time driven by certain rules in high-dimensional space.The dynamic system can be modeled by reconstructing the states that evolve over time.Based on empirical dynamic modeling theory,the causal relationship betwen variables can be detected through the time series data of variables in dynamic system.If variable X is the cause of variable Y(X⇒Y),the information of variable X must be implicit in variable Y and can be recovered from variable Y.This paper first analyzes the dialectical relationship between correlation and causality.Correlation does not mean causality,and lack of correlation does not mean no causality.Then it comprehensively introduces the core idea of causality detection based on empirical dynamic modeling theory,andsummarizes the historical development of Takens embedding theorem,simplex projection algorithm and convergent cross mapping algorithm.It introduces some improved methods of empirical dynamic modeling theory and practical application of causal detection,and finaly looks forward to the future development trend of causal detection based on empirical dynamic modeling.

Key words: Causal detection, Empirical dynamic modeling, Convergent cross mapping, Complex system, Artificia linteligence

中图分类号:

TP391

曹旨昊, 牟少敏, 屈洪春. 基于经验动态建模的因果检测研究综述[J]. 计算机科学, 2023, 50(6A): 220600194-12. https://doi.org/10.11896/jsjkx.220600194

CAO Zhihao, MU Shaomin, QU Hongchun. Review on Causality Detection Based on Empirical Dynamic Modeling[J]. Computer Science, 2023, 50(6A): 220600194-12. https://doi.org/10.11896/jsjkx.220600194

参考文献

[1]ADLER J,PARMRYD I.Quantifying colocalization by correlation:the Pearson correlation coefficient is superior to the Mander’s overlap coefficient[J].Cytometry Part A,2010,77(8):733-742.
[2]SUGIHARA G,MAY R,YE H,et al.Detecting causality incomplex ecosystems[J].science,2012,338(6106):496-500.
[3]ALTMAN N,KRZYWINSKI M.Points of Significance:Asso-ciation,correlation and causation[J].Nature Methods,2015,12(10):899-900.
[4]HUANG Y,FU Z,FRANZKE C L E.Detecting causality from time series in a machine learning framework[J].Chaos:An Interdisciplinary Journal of Nonlinear Science,2020,30(6):063116.
[5]PEARL J,MACKENZIE D.The book of why:the new science of cause and effect[M].Basic Books,2018.
[6]CAO Z,MU S,XU L,et al.Causal Research on Soil Temperature and Moisture Content at Different Depths[J].IEEE Access,2021,9:39077-39088.
[7]PEARL J.Theoretical impediments to machine learning withseven sparks from the causal revolution[J].arXiv:1801.04016,2018.
[8]GRANGER C W J.Investigating causal relations by econometric models and cross-spectral methods[J].Econometrica:Journal of the Econometric Society,1969:424-438.
[9]PEARL J.Causal inference in statistics:An overview[J].Statistics surveys,2009,3:96-146.
[10]LINDEGREN M,CHECKLEY D M,ROUYER T,et al.Climate,fishing,and fluctuations of sardine and anchovy in the California Current[J].Proceedings of the National Academy of Sciences,2013,110(33):13672-13677.
[11]PARK J,PAO G M,SUGIHARA G,et al.Empirical mode mo-deling[J].Nonlinear Dynamics,2022:108(3):2147-2160.
[12]YE H,BEAMISH R J,GLASER S M,et al.Equation-freemechanistic ecosystem forecasting using empirical dynamic mo-deling[J].Proceedings of the National Academy of Sciences,2015,112(13):E1569-E1576.
[13]PAGANO R R.Understanding statistics in the behavioralsciences[M].Cengage Learning,2012.
[14]SIMON H A.Spurious correlation:A causal interpretation[J].Journal of the American statistical Association,1954,49(267):467-479.
[15]GUO F,TIAN L W.Correlation and causality[J].NationalMedical Journal of China,2019(10):790-795.
[16]MA H,LENG S,CHEN L.Data-based prediction and causality inference of nonlinear dynamics[J].Science China Mathematics,2018,61(3):403-420.
[17]WARD A.Spurious correlations and causal inferences[J].Erkenntnis,2013,78(3):699-712.
[18]GLYMOUR M,PEARL J,JEWELL N P.Causal inference in statistics:A primer[M].John Wiley & Sons,2016.
[19]MORAN P A P.The statistical analysis of the Canadian lynxcycle[J].Australian Journal of Zoology,1953,1(3):291-298.
[20]PANCZAK R,GALOBARDES B,SPOERRI A,et al.High life in the sky? Mortality by floor of residence in Switzerland[J].European Journal of Epidemiology,2013,28(6):453-462.
[21]LORENZ E N.Deterministic nonperiodic flow[J].Journal of Atmospheric Sciences,1963,20(2):130-141.
[22]DEANGELIS D L,YUREK S.Equation-free modeling unravels the behavior of complex ecological systems[J].Proceedings of the National Academy of Sciences,2015,112(13):3856-3857.
[23]TAKENS F.Detecting strange attractors in turbulence[M]//Dynamical Systems and Turbulence,Warwick 1980.Springer,1981:366-381.
[24]DEYLE E R,SUGIHARA G.Generalized theorems for nonli-near state space reconstruction[J].Plos One,2011,6(3):e18295.
[25]KANTZ H,SCHREIBER T.Nonlinear time series analysis[M].Cambridge University Press,2004.
[26]BRUNTON S L,PROCTOR J L,KUTZ J N.Discovering governing equations from data by sparse identification of nonlinear dynamical systems[J].Proceedings of the National Academy of Sciences,2016,113(15):3932-3937.
[27]STROGATZ S H.Nonlinear dynamics and chaos:with applications to physics,biology,chemistry,and engineering[M].CRC press,2018.
[28]ASCIOTI F A,BELTRAMI E,CARROLL T O,et al.Is there chaos in plankton dynamics?[J].Journal of Plankton Research,1993,15(6):603-617.
[29]PACKARD N H,CRUTCHFIELD J P,FARMER J D,et al.Geometry from a time series[J].Physical Review Letters,1980,45(9):712.
[30]LIEBERT W,SCHUSTER H G.Proper choice of the time delay for the analysis of chaotic time series[J].Physics Letters A,1989,142(2/3):107-111.
[31]PECORA L M,MONIZ L,NICHOLS J,et al.A unified ap-proach to attractor reconstruction[J].Chaos:An Interdisciplinary Journal of Nonlinear Science,2007,17(1):013110.
[32]KENNEL M B,BROWN R,ABARBANEL H D I.Determining embedding dimension for phase-space reconstruction using a geometrical construction[J].Physical Review A,1992,45(6):3403.
[33]SUGIHARA G,MAY R M.Nonlinear forecasting as a way of distinguishing chaos from measurement error in time series[J].Nature,1990,344(6268):734-741.
[34]FARMER J D,SIDOROWICH J J.Predicting chaotic time series[J].Physical Review Letters,1987,59(8):845.
[35]WHITNEY H.Differentiable manifolds[J].Annals of Mathe-matics,1936:645-680.
[36]CHANG C W,USHIO M,HSIEH C.Empirical dynamic mode-ling for beginners[J].Ecological Research,2017,32(6):785-796.
[37]HSIEH C,GLASER S M,LUCAS A J,et al.Distinguishing random environmental fluctuations from ecological catastrophes for the North Pacific Ocean[J].Nature,2005,435(7040):336-340.
[38]DIXON P A,MILICICH M J,SUGIHARA G.Episodic fluctuations in larval supply[J].Science,1999,283(5407):1528-1530.
[39]YE H,SUGIHARA G.Information leverage in interconnectedecosystems:Overcoming the curse of dimensionality[J].Scien-ce,2016,353(6302):922-925.
[40]SUGIHARA G.Nonlinear forecasting for the classification ofnatural time series[J].Philosophical Transactions of the Royal Society of London.Series A:Physical and Engineering Sciences,1994,348(1688):477-495.
[41]KRAKOVSKÁ A,JAKUBÍK J.Implementation of two causal methods based on predictions in reconstructed state spaces[J].Physical Review E,2020,102(2):022203.
[42]MA H,LENG S,AIHARA K,et al.Randomly distributed embedding making short-term high-dimensional data predictable[J].Proceedings of the National Academy of Sciences,2018,115(43):E9994-E10002.
[43]CUMMINS B,GEDEON T,SPENDLOVE K.On the efficacy of state space reconstruction methods in determining causality[J].SIAM Journal on Applied Dynamical Systems,2015,14(1):335-381.
[44]ARNOLD A,LIU Y,ABE N.Temporal causal modeling with graphical granger methods[C]//Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2007:66-75.
[45]LE VAN QUYEN M,MARTINERIE J,ADAM C,et al.Nonlinear analyses of interictal EEG map the brain interdependences in human focal epilepsy[J].Physica D:Nonlinear Phenomena,1999,127(3/4):250-266.
[46]DAVID O,GUILLEMAIN I,SAILLET S,et al.Identifying neural drivers with functional MRI:an electrophysiological validation[J].PLoS biology,2008,6(12):e315.
[47]HLAVÁČKOVÁ-SCHINDLER K,PALUŠ M,VEJMELKA M,et al.Causality detection based on information-theoretic approaches in time series analysis[J].Physics Reports,2007,441(1):1-46.
[48]PALUŠ M,KRAKOVSKÁ A,JAKUBÍK J,et al.Causality,dynamical systems and the arrow of time[J].Chaos:An Interdisciplinary Journal of Nonlinear Science,2018,28(7):075307.
[49]MCCRACKEN J M,WEIGEL R S.Convergent cross-mapping and pairwise asymmetric inference[J].Physical Review E,2014,90(6):062903.
[50]HSIEH C,OHMAN M D.Biological responses to environmental forcing:the linear tracking window hypothesis[J].Ecology,2006,87(8):1932-1938.
[51]MØNSTER D,FUSAROLI R,TYLÉN K,et al.Causal infe-rence from noisy time-series data－testing the convergent cross-mapping algorithm in the presence of noise and external in-fluence[J].Future Generation Computer Systems,2017,73:52-62.
[52]SCHIFF S J,SO P,CHANG T,et al.Detecting dynamical interdependence and generalized synchrony through mutual prediction in a neural ensemble[J].Physical Review E,1996,54(6):6708.
[53]COUFAL D,JAKUBÍK J,JAJCAY N,et al.Detection of coupling delay:A problem not yet solved[J].Chaos:An Interdisciplinary Journal of Nonlinear Science,2017,27(8):083109.
[54]YE H,DEYLE E R,GILARRANZ L J,et al.Distinguishingtime-delayed causal interactions using convergent cross mapping[J].Scientific Reports,2015,5(1):1-9.
[55]GE X,LIN A.Dynamic causality analysis using overlapped sli-ding windows based on the extended convergent cross-mapping[J].Nonlinear Dynamics,2021,104(2):1753-1765.
[56]MA H,AIHARA K,CHEN L.Detecting causality from nonli-near dynamics with short-term time series[J].Scientific Reports,2014,4(1):1-10.
[57]MA H,LENG S,TAO C,et al.Detection of time delays and directional interactions based on time series from complex dynamical systems[J].Physical Review E,2017,96(1):012221.
[58]HARNACK D,LAMINSKI E,SCH?NEMANN M,et al.Topological causality in dynamical systems[J].Physical Review Letters,2017,119(9):098301.
[59]WANG Y,HU F,CAO Y,et al.Improved CCM for variablecausality detection in complex systems[J].Control Engineering Practice,2019,83:67-82.
[60]DÍAZ E,ADSUARA J E,MARTÍNEZ Á M,et al.Inferringcausal relations from observational long-term carbon and water fluxes records[J].Scientific Reports,2022,12(1):1-12.
[61]CLARK A T,YE H,ISBELL F,et al.Spatial convergent cross mapping to detect causal relationships from short time series[J].Ecology,2015,96(5):1174-1181.
[62]KRAKOVSKÁ A,HANZELY F.Testing for causality in reconstructed state spaces by an optimized mixed prediction method[J].Physical Review E,2016,94(5):052203.
[63]LENG S,MA H,KURTHS J,et al.Partial cross mapping eliminates indirect causal influences[J].Nature communications,2020,11(1):1-9.
[64]JOHNSON B,MUNCH S B.An empirical dynamic modelingframework for missing or irregular samples[J].Ecological Mo-delling,2022,468:109948.
[65]JIANG J J,HUANG Z G,HUANG L,et al.Directed dynamical influence is more detectable with noise[J].Scientific Reports,2016,6(1):1-9.
[66]KAWATSU K,USHIO M,VAN VEEN F J F,et al.Are networks of trophic interactions sufficient for understanding the dynamics of multi-trophic communities? Analysis of a tri-trophic insect food-web time-series[J].Ecology Letters,2021,24(3):543-552.
[67]MCGOWAN J A,DEYLE E R,YE H,et al.Predicting coastal algal blooms in southern California[J].Ecology,2017,98(5):1419-1433.
[68]YE L,TAN L,WU X,et al.Nonlinear causal analysis reveals an effective water level regulation approach for phytoplankton blooms controlling in reservoirs[J].Science of The Total Environment,2022,806:150948.
[69]USHIO M,HSIEH C,MASUDA R,et al.Fluctuating interaction network and time-varying stability of a natural fish community[J].Nature,2018,554(7692):360-363.
[70]CHANG C W,MIKI T,YE H,et al.Causal networks of phytoplankton diversity and biomass are modulated by environmental context[J].Nature Communications,2022,13(1):1-11.
[71]KITAYAMA K,USHIO M,AIBA S I.Temperature is a dominant driver of distinct annual seasonality of leaf litter production of equatorial tropical rain forests[J].Journal of Ecology,2021,109(2):727-736.
[72]SHI H,ZHAO Y,LIU S,et al.A new perspective on drought propagation:Causality[J].Geophysical Research Letters,2022,49(2):e2021GL096758.
[73]DOI H,YASUHARA M,USHIO M.Causal analysis of thetemperature impact on deep-sea biodiversity[J].Biology Letters,2021,17(7):20200666.
[74]WEN C,YANG J.Complexity evolution of chaotic financial systems based on fractional calculus[J].Chaos,Solitons & Fractals,2019,128:242-251.
[75]WANG Z,HUANG X,SHI G.Analysis of nonlinear dynamics and chaos in a fractional order financial system with time delay[J].Computers & Mathematics with Applications,2011,62(3):1531-1539.
[76]HUFFAKER R,FEARNE A.Empirically testing for dynamiccausality between promotions and sales beer promotions and sales in england[C]//2014 International European Forum.2014:17-21.
[77]AZQUETA-GAVALDN A.Causal inference between cryp-tocurrency narratives and prices:Evidence from a complex dynamic ecosystem[J].Physica A:Statistical Mechanics and its Applications,2020,537:122574.
[78]VISSIO G,LUCARINI V.A proof of concept for scale-adaptive parametrizations:the case of the Lorenz’96 model[J].Quarterly Journal of the Royal Meteorological Society,2018,144(710):63-75.
[79]TSONIS A A,DEYLE E R,MAY R M,et al.Dynamical evidence for causality between galactic cosmic rays and interannual variation in global temperature[J].Proceedings of the National Academy of Sciences,2015,112(11):3253-3256.
[80]WANG X,PIAO S,CIAIS P,et al.A two-fold increase of carbon cycle sensitivity to tropical temperature variations[J].Nature,2014,506(7487):212-215.
[81]VAN NES E H,SCHEFFER M,BROVKIN V,et al.Causalfeedbacks in climate change[J].Nature Climate Change,2015,5(5):445-448.
[82]WANG Y,YANG J,CHEN Y,et al.Detecting the causal effect of soil moisture on precipitation using convergent cross mapping[J].Scientific Reports,2018,8(1):1-8.
[83]HUANG Y,FRANZKE C L E,YUAN N,et al.Systematicidentification of causal relations in high-dimensional chaotic systems:application to stratosphere-troposphere coupling[J].Climate Dynamics,2020,55(9):2469-2481.
[84]ZHANG N,WANG G.Detecting the causal interaction between Siberian high and winter surface air temperature over northeast Asia[J].Atmospheric Research,2020,245:105066.
[85]HESKAMP L,MEEL-VAN DEN ABEELEN A S,LAGRO J,et al.Convergent cross mapping:a promising technique for cerebral autoregulation estimation[J].Int.J.Clin.Neurosci.Ment.Heal.S,2014,20.
[86]MCBRIDE J C,ZHAO X,MUNRO N B,et al.Sugihara causality analysis of scalp EEG for detection of early Alzheimer’s disease[J].NeuroImage:Clinical,2015,7:258-265.
[87]DEYLE E R,MAHER M C,HERNANDEZ R D,et al.Global environmental drivers of influenza[J].Proceedings of the National Academy of Sciences,2016,113(46):13081-13086.
[88]MÜLLER A,KRAEMER J F,PENZEL T,et al.Causality inphysiological signals[J].Physiological Measurement,2016,37(5):R46.
[89]VERMA A K,GARG A,XU D,et al.Skeletal muscle pump drives control of cardiovascular and postural systems[J].Scientific Reports,2017,7(1):1-8.
[90]COBEY S,BASKERVILLE E B.Limits to causal inference with state-space reconstruction for infectious disease[J].PloS One,2016,11(12):e0169050.
[91]BOZORGMAGHAM A E,MOTESHARREI S,PENNY S G,et al.Causality analysis:identifying the leading element in a coupled dynamical system[J].PLoS One,2015,10(6):e0131226.
[92]LUO L,CHENG F,QIU T,et al.Refined convergent cross-mapping for disturbance propagation analysis of chemical processes[J].Computers & Chemical Engineering,2017,106:1-16.
[93]XIANG S,BAI Y,ZHAO J.Medium-term prediction of keychemical process parameter trend with small data[J].Chemical Engineering Science,2022,249:117361.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed