Pre-trained Graph Learning Model with Spatial-aware Attention Mechanism for Road Network Patterns Recognition

This repository contains the official implementation for the paper: "Pre-trained Graph Learning Model with Spatial-aware Attention Mechanism for Road Network Patterns Recognition". It aims to classify the input road networks into six road network patterns: Gridiron, Organic, Radial, Tributary, Linear, and Chaotic. We represent road networks as graphs and employ a pre-training strategy on a large corpus of unlabeled road network data. Our work introduces a novel pre-trained graph learning model that leverages a spatial-aware attention mechanism that allows the model to focus on salient geometric and topological features, leading to superior performance in pattern recognition tasks.