Second-hand house price index forecasting with neural networks

Abstract

ABSTRACT The house market in China has been growing rapidly for the past decade and price forecasting has become a significant issue to the people and policymakers. We approach this problem by examining neural networks for second-hand house price index forecasting from 10 major cities for March 2012–May 2020. Our purpose is to build simple and accurate neural networks to contribute to pure technical house price forecasting of the Chinese market. We explore various model settings across the algorithm, delay, hidden neuron, and data spitting ratio, and arrive at a rather simple neural network with three delays and eight hidden neurons, which leads to stable performance of 0.8% average relative root-mean-square error across the 10 cities for the training, validation, and testing phases. Results here can be used on a standalone basis or combined with fundamental forecasting in forming perspectives of house price trends and conducting policy analysis.