![[2602.21381] VCDF: A Validated Consensus-Driven Framework for Time Series Causal Discovery](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.21381] VCDF: A Validated Consensus-Driven Framework for Time Series Causal Discovery
Summary
The paper presents VCDF, a framework designed to enhance the robustness of time series causal discovery methods by evaluating the stability of causal relationships across temporal subsets.
Why It Matters
Time series causal discovery is crucial for understanding dynamic systems in various fields, including finance and neuroscience. VCDF addresses the limitations of existing methods, improving their reliability in the presence of noise and variability, which is essential for accurate data interpretation and decision-making.
Key Takeaways
- VCDF improves the stability of causal relations in time series analysis.
- The framework is method-agnostic and can enhance existing algorithms without modification.
- Significant improvements in performance metrics were observed, especially with longer data sequences.
- VCDF demonstrates robustness under realistic noise conditions, making it applicable in practical scenarios.
- The framework can be beneficial for researchers and practitioners in dynamic system analysis.
Computer Science > Machine Learning arXiv:2602.21381 (cs) [Submitted on 22 Feb 2026] Title:VCDF: A Validated Consensus-Driven Framework for Time Series Causal Discovery Authors:Gene Yu, Ce Guo, Wayne Luk View a PDF of the paper titled VCDF: A Validated Consensus-Driven Framework for Time Series Causal Discovery, by Gene Yu and 2 other authors View PDF HTML (experimental) Abstract:Time series causal discovery is essential for understanding dynamic systems, yet many existing methods remain sensitive to noise, non-stationarity, and sampling variability. We propose the Validated Consensus-Driven Framework (VCDF), a simple and method-agnostic layer that improves robustness by evaluating the stability of causal relations across blocked temporal subsets. VCDF requires no modification to base algorithms and can be applied to methods such as VAR-LiNGAM and PCMCI. Experiments on synthetic datasets show that VCDF improves VAR-LiNGAM by approximately 0.08-0.12 in both window and summary F1 scores across diverse data characteristics, with gains most pronounced for moderate-to-long sequences. The framework also benefits from longer sequences, yielding up to 0.18 absolute improvement on time series of length 1000 and above. Evaluations on simulated fMRI data and IT-monitoring scenarios further demonstrate enhanced stability and structural accuracy under realistic noise conditions. VCDF provides an effective reliability layer for time series causal discovery without altering underlying mo...
