Causal Lag Structure Discovery in Confounded Time Series via Orthogonalized Adaptive Estimation

研究问题: 多元时间序列存在非线性混杂和时变结构时，如何同时调整混杂、选择滞后阶数，并对定向关系的发现进行错误率控制？
方法与思路: 作者提出 ORACLE-VARX：用 DML 调整混杂，以序贯检验选择滞后，再对系数进行检验和多重校正，并在滚动窗口中分析。
材料陈述的贡献: 摘要称，窗口平均目标的去偏系数渐近正态；特定模拟显示滞后追踪与预测改进，并在美国行业 ETF 应用中得到因果图。
阅读关注点: 应核对因果解释所需条件、窗口平均目标、时序依赖与滞后选择的处理，以及多重校正能否在这些步骤后实现所称控制。
阅读边界: 这是 arXiv v1 预印本，不能据此声称已同行评审或期刊接收；摘要未说明完整因果识别及错误率控制条件，模拟数值不足以验证应用中的因果解释。

GPT-6.1 Sol · 高 · 2026-10-10

原文摘要

Finding which variables cause which others in multivariate time series, and at what lags, is central to science and policy, yet existing methods force a choice between flexible confounder adjustment, data-driven lag selection, and inference that controls the false discovery rate (FDR). ORACLE-VARX does all three in one pipeline. First, double/debiased machine learning (DML) removes nonlinear confounder effects from the outcomes and the lagged series. Second, adaptive causal lag estimation (ACLE) picks the lag order at each time step by sequential significance tests, tracking regime changes. Third, entry-wise $z$ -tests with Benjamini--Hochberg correction select directed edges at a target FDR. We prove that in each rolling window, the debiased coefficients are asymptotically normal around a window-averaged target, so their $z$ -tests are asymptotically valid. On a synthetic benchmark with time-varying structure and nonlinear confounding, ORACLE-VARX (LightGBM) tracks the true lag order best (RMSE $0.96$ vs $1.1$ -- $1.5$ ), has edge FDR $0.047$ , close to PCMCI ( $0.045$ ) and below VAR ( $0.129$ ) and VAR-LiNGAM ( $0.187$ ), and forecasts better than all three. On nine U.S. sector ETFs with macroeconomic confounders, it yields interpretable causal graphs whose lag order rises in high-volatility regimes.Finding which variables cause which others in multivariate time series, and at what lags, is central to science and policy, yet existing methods force a choice between flexible confounder adjustment, data-driven lag selection, and inference that controls the false discovery rate (FDR). ORACLE-VARX does all three in one pipeline. First, double/debiased machine learning (DML) removes nonlinear confounder effects from the outcomes and the lagged series. Second, adaptive causal lag estimation (ACLE) picks the lag order at each time step by sequential significance tests, tracking regime changes. Third, entry-wise $z$-tests with Benjamini--Hochberg correction select directed edges at a target FDR. We prove that in each rolling window, the debiased coefficients are asymptotically normal around a window-averaged target, so their $z$-tests are asymptotically valid. On a synthetic benchmark with time-varying structure and nonlinear confounding, ORACLE-VARX (LightGBM) tracks the true lag order best (RMSE $0.96$ vs $1.1$--$1.5$), has edge FDR $0.047$, close to PCMCI ($0.045$) and below VAR ($0.129$) and VAR-LiNGAM ($0.187$), and forecasts better than all three. On nine U.S. sector ETFs with macroeconomic confounders, it yields interpretable causal graphs whose lag order rises in high-volatility regimes.

摘要来源：arXiv

笔记与分组

同方向的基础文献

相关领域论文

共同方法：因果推断、时间序列、高维与机器学习、检验与推断、一般估计、VAR 与局部投影、预测与混频、机器学习与 DML · 共同方向：时间序列、VAR 与局部投影

Semiparametric Inference for Dynamic Causal Effects from Observational Time Series

Shibo Yu · Yan Chen · Jin-Hong Du · Guodong Li首次提交 2026-09-23 · 修订 2026-09-23 · v1arXiv预印本

观察性时间序列动态因果效应的半参数推断AI 解读

In observational time series, statistical inference for dynamic causal effects of a one-time intervention across horizons is complicated by high-dimensional observed pre-treatment information, unmeasured confounding, and serial dependence. To address these challenges, we develop a semiparametric framework for inference from a single serially dependent time series, integrating debiased machine learning with instrumental variables through buffered block cross-fitting. Under…

微观 · 因果推断半参数与效率预测与混频

关联记录被引 0 次 ↗OpenAlex · 核验 2026-10-01

AI 解读阅读原文 ↗BibTeX ↓

共同方法：时间序列、结构变化与时变参数、高维与机器学习、检验与推断、一般估计、VAR 与局部投影、预测与混频、机器学习与 DML · 共同方向：时间序列、VAR 与局部投影

Interpreting and predicting the economy flows: A time-varying parameter global vector autoregressive integrated the machine learning model

Yukang Jiang · Xueqin Wang · Zhixi Xiong · Haisheng Yang · Ting Tian首次提交 2022-07-31 · 修订 2022-07-31 · v1arXiv预印本

融合机器学习的时变参数全球向量自回归：经济动态解释与预测AI 解读

The paper proposes a time-varying parameter global vector autoregressive (TVP-GVAR) framework for predicting and analysing developed region economic variables. We want to provide an easily accessible approach for the economy application settings, where a variety of machine learning models can be incorporated for out-of-sample prediction. The LASSO-type technique for numerically efficient model selection of mean squared errors (MSEs) is selected. We show the convincing…

宏观 · 时间序列结构变化与时变参数机器学习与 DML

关联记录被引 0 次 ↗OpenAlex · 核验 2026-09-30

AI 解读阅读原文 ↗BibTeX ↓

共同方法：因果推断、时间序列、高维与机器学习、检验与推断、一般估计、VAR 与局部投影、机器学习与 DML · 共同方向：时间序列、VAR 与局部投影

Double machine learning for time series

Milos Ciganovic · Federico D’Amario · Massimiliano Tancioni上线 2026-07-06The Econometrics Journal在线待刊

Summary We adapt double machine learning to macroeconomic time series by combining regularized nuisance estimation with reverse cross-fitting. This deterministic scheme exploits time reversibility to use time-reversed auxiliary blocks and, unlike neighbour-deletion designs, avoids buffer blocks, thereby improving sample usage. We derive conditions for asymptotic validity and show in simulations that the estimator performs well in realistic finite samples across the designs…

宏观 · 时间序列因果推断机器学习与 DML

被引 0 次 ↗OpenAlex · 核验 2026-10-15

阅读原文 ↗BibTeX ↓

共同方法：时间序列、结构变化与时变参数、高维与机器学习、检验与推断、一般估计、VAR 与局部投影、预测与混频 · 共同方向：时间序列、VAR 与局部投影

Time-Varying Model Averaging of Multi-layer Network Vector Autoregressions

Degui Li · Yuying Sun · Boyao Wu首次提交 2026-06-24 · 修订 2026-06-24 · v1arXiv预印本

多层网络向量自回归的时变模型平均AI 解读

In this paper, we introduce a flexible time-varying multi-layer network vector autoregression (VAR) model framework for large-scale time series, allowing agents in dynamic systems to interact through multiple channels and incorporating multiple adjacency matrices to capture network spillover effects. We propose a penalized model averaging method to determine a time-varying optimal combination of multi-layer network VAR candidate models whose number may be divergent. Under…

宏观 · 时间序列结构变化与时变参数空间与网络

关联记录被引 0 次 ↗OpenAlex · 核验 2026-10-01

AI 解读阅读原文 ↗BibTeX ↓

共同方法：因果推断、时间序列、高维与机器学习、检验与推断、一般估计、VAR 与局部投影、机器学习与 DML · 共同方向：时间序列、VAR 与局部投影

Semiparametric inference for impulse response functions using double/debiased machine learning

Daniele Ballinari · Alexander Wehrli首次提交 2024-11-15 · 修订 2025-12-16 · v2arXiv预印本

用双重／去偏机器学习对脉冲响应函数进行半参数推断AI 解读

We introduce a double/debiased machine learning estimator for the impulse response function in settings where a time series of interest is subjected to multiple discrete treatments, assigned over time, which can have a causal effect on future outcomes. The proposed estimator can rely on fully nonparametric relations between treatment and outcome variables, opening up the possibility to use flexible machine learning approaches to estimate impulse response functions. To this…

微观 · 因果推断半参数与效率机器学习与 DML

关联记录被引 0 次 ↗OpenAlex · 核验 2026-09-30

AI 解读阅读原文 ↗BibTeX ↓

共同方法：时间序列、高维与机器学习、检验与推断、一般估计、VAR 与局部投影、预测与混频、多重检验与同时推断 · 共同方向：时间序列、VAR 与局部投影

Uniform Inference in High-Dimensional Threshold Regression Models

Jiatong Li · Hongqiang Yan首次提交 2024-04-11 · 修订 2025-09-14 · v3arXiv预印本

高维门槛回归模型中的一致推断AI 解读

We develop a uniform inference theory for high-dimensional slope parameters in threshold regression models, allowing for either cross-sectional or time series data. We first establish oracle inequalities for prediction errors, and L1 estimation errors for the Lasso estimator of the slope parameters and the threshold parameter, accommodating heteroskedastic non-subgaussian error terms and non-subgaussian covariates. Next, we derive the asymptotic distribution of tests…

宏观 · 时间序列多重检验与同时推断预测与混频

关联记录被引 0 次 ↗OpenAlex · 核验 2026-09-30

AI 解读阅读原文 ↗BibTeX ↓