---
title: "Causal Discovery on Irregular Time Series"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.18226v1"
permalink: "https://feed.astralor.com/items/itm_047ae2d4472643e1c7919de6"
published: "2026-07-20T17:57:45.000Z"
score: 72
channel: "ai"
tags: ["PCMCI+","Irregular Time Series","Causal Discovery","Window Aggregation"]
---

# Causal Discovery on Irregular Time Series

## AI commentary

> This paper is a valuable addition to causal discovery, especially for readers working with sensor streams, healthcare time series, or financial transaction data. The extension of PCMCI+ is straightforward yet effective, and the solid synthetic results make it worth following up with real-world applications.

## At a glance

This paper proposes an extension of PCMCI+ for causal discovery on irregularly sampled time series. Traditional methods rely on fixed-lag structures, limiting applicability to irregular streams such as sensor data, healthcare records, and financial transactions. The new approach aggregates causal influences over predefined temporal windows, consistently recovering the true causal graph and substantially outperforming standard PCMCI+ on synthetic irregular data.

## Key points

- Traditional causal discovery relies on regular discrete lags, failing on irregular time series
- Proposes an extension of PCMCI+ that aggregates causal influences over temporal windows
- Outperforms standard PCMCI+ on synthetic irregular event streams
- Applicable to sensor streams, healthcare data, and financial transactions

## Full text

Causal discovery methods have shown strong performance in temporal systems, but they typically rely on regular and discrete lag structures, limiting their applicability to regularly sampled data. However, many real-world tasks require dealing with irregularly sampled streams of events, such as sensor streams, healthcare data, and financial transactions. In this work, we propose an extension of PCMCI+, a state-of-the-art method for causal discovery on regular multivariate time series, to allow for handling irregular time series. Instead of modelling causal relations through fixed-lag dependencies, our method aggregates causal influence over predefined temporal windows. We evaluate our method on synthetic irregular event streams with known causal structures under different signal-to-noise ratios, showing that it consistently recovers the underlying causal graph and substantially outperforms the standard PCMCI+ on irregularly sampled data.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_047ae2d4472643e1c7919de6