Event-CausNet: Unlocking Causal Knowledge from Text with Large Language Models for Reliable Spatio-Temporal Forecasting

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Niu, Luyao, Wang, Zepu, Guan, Shuyi, Liu, Yang, Sun, Peng
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912713070870528
author Niu, Luyao
Wang, Zepu
Guan, Shuyi
Liu, Yang
Sun, Peng
author_facet Niu, Luyao
Wang, Zepu
Guan, Shuyi
Liu, Yang
Sun, Peng
contents While spatio-temporal Graph Neural Networks (GNNs) excel at modeling recurring traffic patterns, their reliability plummets during non-recurring events like accidents. This failure occurs because GNNs are fundamentally correlational models, learning historical patterns that are invalidated by the new causal factors introduced during disruptions. To address this, we propose Event-CausNet, a framework that uses a Large Language Model to quantify unstructured event reports, builds a causal knowledge base by estimating average treatment effects, and injects this knowledge into a dual-stream GNN-LSTM network using a novel causal attention mechanism to adjust and enhance the forecast. Experiments on a real-world dataset demonstrate that Event-CausNet achieves robust performance, reducing prediction error (MAE) by up to 35.87%, significantly outperforming state-of-the-art baselines. Our framework bridges the gap between correlational models and causal reasoning, providing a solution that is more accurate and transferable, while also offering crucial interpretability, providing a more reliable foundation for real-world traffic management during critical disruptions.
format Preprint
id arxiv_https___arxiv_org_abs_2511_12769
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Event-CausNet: Unlocking Causal Knowledge from Text with Large Language Models for Reliable Spatio-Temporal Forecasting
Niu, Luyao
Wang, Zepu
Guan, Shuyi
Liu, Yang
Sun, Peng
Artificial Intelligence
Machine Learning
While spatio-temporal Graph Neural Networks (GNNs) excel at modeling recurring traffic patterns, their reliability plummets during non-recurring events like accidents. This failure occurs because GNNs are fundamentally correlational models, learning historical patterns that are invalidated by the new causal factors introduced during disruptions. To address this, we propose Event-CausNet, a framework that uses a Large Language Model to quantify unstructured event reports, builds a causal knowledge base by estimating average treatment effects, and injects this knowledge into a dual-stream GNN-LSTM network using a novel causal attention mechanism to adjust and enhance the forecast. Experiments on a real-world dataset demonstrate that Event-CausNet achieves robust performance, reducing prediction error (MAE) by up to 35.87%, significantly outperforming state-of-the-art baselines. Our framework bridges the gap between correlational models and causal reasoning, providing a solution that is more accurate and transferable, while also offering crucial interpretability, providing a more reliable foundation for real-world traffic management during critical disruptions.
title Event-CausNet: Unlocking Causal Knowledge from Text with Large Language Models for Reliable Spatio-Temporal Forecasting
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2511.12769