DCD: Decomposition-based Causal Discovery from Autocorrelated and Non-Stationary Temporal Data

2026-02-01

AI Summary

This work addresses the challenge of spurious associations and misidentified temporal dependencies in causal inference from multivariate time series, which often arise due to non-stationarity and autocorrelation. To this end, the authors propose a multiscale causal discovery method that first decomposes each time series into trend, seasonal, and residual components. These components are then analyzed separately using stationarity testing, kernel-based dependence measures, and constraint-based causal discovery, respectively—enabling, for the first time, the disentanglement of long-term and short-term causal effects. The results from each component are subsequently integrated to construct a unified causal graph. Empirical evaluations on both synthetic and real-world climate data demonstrate that the proposed approach significantly reduces spurious links and outperforms existing methods, particularly under strong non-stationarity and high autocorrelation, thereby recovering more accurate causal structures and enhancing model interpretability and performance.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

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.

Problem

• causal discovery
• non-stationary time series
• autocorrelation
• spurious causality
• temporal data

Innovation

• decomposition-based causal discovery
• non-stationary time series
• autocorrelated data
• multi-scale causal structure
• spurious edge reduction