Conformal Load Prediction with Transductive Graph Autoencoders

TL;DR

This paper introduces a conformal prediction method combined with graph neural networks for accurate and reliable edge weight prediction in graphs, ensuring valid prediction intervals with improved coverage and efficiency.

Contribution

It presents a novel conformal load prediction framework using transductive graph autoencoders that guarantees coverage and handles data heteroscedasticity effectively.

Findings

Outperforms baseline methods in coverage and efficiency

Demonstrates robustness across transportation datasets

Provides valid prediction intervals with guaranteed coverage

Abstract

Predicting edge weights on graphs has various applications, from transportation systems to social networks. This paper describes a Graph Neural Network (GNN) approach for edge weight prediction with guaranteed coverage. We leverage conformal prediction to calibrate the GNN outputs and produce valid prediction intervals. We handle data heteroscedasticity through error reweighting and Conformalized Quantile Regression (CQR). We compare the performance of our method against baseline techniques on real-world transportation datasets. Our approach has better coverage and efficiency than all baselines and showcases robustness and adaptability.