方法对比

	机器学习增强型中断时间序列	动态中断时间序列
领域	因果推断	因果推断
方法族	Regression model	Regression model
起源年份	2014-2015	2002–2017
提出者	Brodersen et al. (2015); Varian (2014) — foundational ML-for-causal-inference literature	Wagner, Soumerai, Zhang & Ross-Degnan; extended by Lopez Bernal, Cummins & Gasparrini
类型	Quasi-experimental causal inference with ML counterfactual	Quasi-experimental time-series design
开创性文献	Brodersen, K. H., Gallusser, F., Koehler, J., Remy, N., & Scott, S. L. (2015). Inferring causal impact using Bayesian structural time-series models. Annals of Applied Statistics, 9(1), 247-274.	Lopez Bernal, J., Cummins, S., & Gasparrini, A. (2017). Interrupted time series regression for the evaluation of public health interventions: a tutorial. International Journal of Epidemiology, 46(1), 348-355.
别名	ML-ITS, ML-augmented ITS, machine learning ITS, causal ML interrupted time series	Dynamic ITS, ITS with lagged effects, time-varying ITS, flexible ITS
相关	6	4
摘要	Machine Learning-Augmented Interrupted Time Series (ML-ITS) estimates the causal effect of a discrete intervention by training a machine learning model on pre-intervention time series data, projecting a counterfactual trajectory into the post-intervention period, and measuring the gap between observed and predicted outcomes. It extends classical ITS by replacing parametric trend assumptions with flexible ML estimators such as gradient boosting, random forests, or Bayesian structural time-series models.	Dynamic Interrupted Time Series (Dynamic ITS) extends the standard ITS design by allowing intervention effects to build up, decay, or shift over multiple time lags rather than assuming a single instantaneous level change. It estimates how an intervention's impact evolves across time periods, making it especially suited to public health, health services research, and policy evaluation where effects accumulate gradually or wear off after initial impact.