Scalable Property Valuation Models via Graph-based Deep Learning

TL;DR

This paper introduces scalable graph neural network models, including transformer-based variants, to improve property valuation accuracy by capturing complex spatial relationships among houses in large datasets.

Contribution

The paper presents two novel graph neural network architectures for property valuation, enhancing scalability and accuracy over existing models.

Findings

Transformer graph convolutions outperform standard methods.

Models significantly improve house price prediction accuracy.

Scalable approach effective on large proprietary dataset.

Abstract

This paper aims to enrich the capabilities of existing deep learning-based automated valuation models through an efficient graph representation of peer dependencies, thus capturing intricate spatial relationships. In particular, we develop two novel graph neural network models that effectively identify sequences of neighboring houses with similar features, employing different message passing algorithms. The first strategy consider standard spatial graph convolutions, while the second one utilizes transformer graph convolutions. This approach confers scalability to the modeling process. The experimental evaluation is conducted using a proprietary dataset comprising approximately 200,000 houses located in Santiago, Chile. We show that employing tailored graph neural networks significantly improves the accuracy of house price prediction, especially when utilizing transformer convolutional message passing layers.