---
title: "Causal-TS：面向高维非平稳时间序列因果发现的开源Python库"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.24673v1"
permalink: "https://feed.astralor.com/items/itm_c47e6aba439748552c21be22"
published: "2026-07-27T17:14:48.000Z"
score: 78
channel: "ai"
tags: ["Bloomberg","Causal-TS","CDNOTS","PyTorch","因果发现","时间序列分析"]
---

# Causal-TS：面向高维非平稳时间序列因果发现的开源Python库

## AI 点评

> 这篇论文值得所有从事时间序列因果推断的研究者关注。它把分散的算法集成到一个GPU加速的统一框架中，并创造性解决了非平稳性的断点问题——这在金融风控和气候归因中是痛点。如果你是数据科学家或ML工程师，直接pip install就能跑起因果图，非常实用。

## 速览

Causal-TS是Bloomberg推出的开源Python库，专为高维非平稳多元时间序列设计。它提供四种专用算法（CDNOTS、CDNOTS+、CEDAR、GRACE）并封装了GES、Granger等经典方法，统一采用PyTorch GPU加速的条件独立性检验层。库中包含区间发现流程，能自动检测结构断点并在各区间内运行因果发现。通过命令行接口、合成数据生成器和可选的DoWhy集成，实现了从原始时间序列到因果效应估计的端到端流程。支持Python 3.10-3.12，pip可安装。

## 关键点

- 提供四种专用因果发现算法及多种经典方法封装，统一CI测试层支持GPU加速。
- 内置区间发现流程，可自动检测时间序列中的结构断点并分区分析。
- 集成DoWhy，实现从原始序列到因果效应估计的完整流程。
- 提供命令行界面与合成数据生成器，降低使用门槛。

## 全文

We describe Causal-TS, an open-source Python library for causal discovery in high-dimensional and nonstationary multivariate time series. Causal-TS provides four specialized algorithms-CDNOTS, CDNOTS+, CEDAR, and GRACE-along with wrappers for GES, Granger, LASSO-VAR, and LGES, all sharing a unified conditional independence (CI) test layer with GPU acceleration via PyTorch. A regime discovery pipeline detects structural breaks via pluggable changepoint detectors and runs discovery per regime with regime-specific parameters. A command-line interface, synthetic data generators, and optional DoWhy integration provide an end-to-end pipeline from raw time series to causal effect estimates. The library is pip-installable, tested on Python 3.10--3.12, and available at https://github.com/bloomberg/causal-ts.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_c47e6aba439748552c21be22