Uncertainty Quantification via Spatial-Temporal Tweedie Model for Zero-inflated and Long-tail Travel Demand Prediction

TL;DR

This paper introduces the Spatial-Temporal Tweedie Graph Neural Network (STTD), a novel model that effectively predicts and quantifies uncertainty in sparse, long-tail origin-destination travel demand matrices using Tweedie distribution.

Contribution

The paper proposes STTD, integrating Tweedie distribution with spatial-temporal embeddings in a graph neural network to improve travel demand prediction and uncertainty quantification.

Findings

STTD outperforms traditional models in accuracy.

STTD provides reliable confidence intervals.

Effective in high-resolution demand scenarios.

Abstract

Understanding Origin-Destination (O-D) travel demand is crucial for transportation management. However, traditional spatial-temporal deep learning models grapple with addressing the sparse and long-tail characteristics in high-resolution O-D matrices and quantifying prediction uncertainty. This dilemma arises from the numerous zeros and over-dispersed demand patterns within these matrices, which challenge the Gaussian assumption inherent to deterministic deep learning models. To address these challenges, we propose a novel approach: the Spatial-Temporal Tweedie Graph Neural Network (STTD). The STTD introduces the Tweedie distribution as a compelling alternative to the traditional 'zero-inflated' model and leverages spatial and temporal embeddings to parameterize travel demand distributions. Our evaluations using real-world datasets highlight STTD's superiority in providing accurate predictions and precise confidence intervals, particularly in high-resolution scenarios.