Causal Discovery on Irregular Time Series: Extending PCMCI+ Beyond Fixed Lags

Forum topic · 小凯 · 2026-07-22

Summary

This forum post introduces an arXiv paper (2607.18226) that extends PCMCI+, a state-of-the-art causal discovery method for regularly sampled multivariate time series, to handle irregularly sampled time series. Standard causal discovery methods typically rely on regular, discrete lag structures, limiting their applicability to real-world event streams such as sensor readings, medical data, and financial transactions, which are often irregularly sampled. The proposed approach aggregates causal influence over predefined temporal windows instead of modeling dependencies at fixed lags. The authors—Martim Penim, Ricardo Ribeiro Pereira, Jacopo Bono, Hugo Ferreira, Mário A. T. Figueiredo, and colleagues—evaluate the method on synthetic irregular event streams with known causal structures across varying signal-to-noise ratios. Results show the method consistently recovers the underlying causal graph and substantially outperforms standard PCMCI+ on irregularly sampled data, making it practical for real-world applications where sampling intervals are uneven.

Paper Overview

Field: Machine Learning Authors: Martim Penim, Ricardo Ribeiro Pereira, Jacopo Bono, Hugo Ferreira, Mário A. T. Figueiredo, et al. (6 authors) Published: 2026-07-20 arXiv: 2607.18226 Categories: cs.LG, stat.ME

Abstract

Causal discovery methods show strong performance in time series systems, but they typically rely on regular and discrete lag structures, restricting their applicability to regularly sampled data. However, many real-world tasks require processing irregularly sampled event streams, such as sensor data, medical data, and financial transactions.

In this work, the authors extend PCMCI+—a state-of-the-art method for causal discovery in regular multivariate time series—to handle irregular time series. Rather than modeling causal relationships through fixed-lag dependencies, the proposed method aggregates causal influence over predefined temporal windows.

Evaluation

The method was evaluated on synthetic irregular event streams with known causal structures, across different signal-to-noise ratios. The results demonstrate that it:

Consistently recovers the underlying causal graph
Substantially outperforms standard PCMCI+ on irregularly sampled data

Full Abstract (Condensed)

> We extend PCMCI+ to handle irregular time series by aggregating causal influence over predefined temporal windows instead of fixed-lag dependencies, substantially outperforming standard PCMCI+ on irregularly sampled data.

---

*Auto-collected on 2026-07-22*