DCD: Decomposition-based Causal Discovery from Autocorrelated and Non-Stationary Temporal Data
This addresses causal inference challenges in domains like finance, climate science, and healthcare, offering improved accuracy under non-stationarity and autocorrelation, but it is incremental as it builds on existing causal discovery methods.
The paper tackled the problem of causal discovery from autocorrelated and non-stationary temporal data by introducing a decomposition-based framework that separates time series into trend, seasonal, and residual components for causal analysis, resulting in more accurate recovery of ground-truth causal structure than state-of-the-art baselines in synthetic benchmarks and real-world climate data.
Multivariate time series in domains such as finance, climate science, and healthcare often exhibit long-term trends, seasonal patterns, and short-term fluctuations, complicating causal inference under non-stationarity and autocorrelation. Existing causal discovery methods typically operate on raw observations, making them vulnerable to spurious edges and misattributed temporal dependencies. We introduce a decomposition-based causal discovery framework that separates each time series into trend, seasonal, and residual components and performs component-specific causal analysis. Trend components are assessed using stationarity tests, seasonal components using kernel-based dependence measures, and residual components using constraint-based causal discovery. The resulting component-level graphs are integrated into a unified multi-scale causal structure. This approach isolates long- and short-range causal effects, reduces spurious associations, and improves interpretability. Across extensive synthetic benchmarks and real-world climate data, our framework more accurately recovers ground-truth causal structure than state-of-the-art baselines, particularly under strong non-stationarity and temporal autocorrelation.