Predicting Second-Hand Housing Prices in Beijing: A Comparative Study of Machine Learning and Ensemble Models

Jia He

2025

Abstract

In recent decades, the real estate market has always been an imperative force boosting China’s economic development. Accurate housing price prediction plays a key role in policymaking and investment decisions. Nonetheless, the traditional models often fall short when dealing with complex and nonlinear relationships. This study focuses on predicting second-hand housing prices in Beijing using machine learning techniques. The web crawler is used to obtain historical transaction data, then yield a dataset that includes 56,793 records with 23 features after pre-processing in Beijing. Three models-Random Forest, XGBoost, and LightGBM- were trained using grid-search and five-fold cross-validation. A combined ensemble model was also built to improve the overall robustness. Evaluation and visualizations were used to compare performance. The ensemble model outperformed the single model, followed by XGBoost, Random Forest, and LighGBM. This paper aims to provide a new idea for house price prediction research through machine learning methods, hoping to bring some inspiration to theoretical research and practical applications in related fields.

Paper Citation

in Harvard Style

He J. (2025). Predicting Second-Hand Housing Prices in Beijing: A Comparative Study of Machine Learning and Ensemble Models. In Proceedings of the 2nd International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI; ISBN 978-989-758-792-4, SciTePress, pages 574-582. DOI: 10.5220/0014363000004718

in Bibtex Style

@conference{emiti25,
author={Jia He},
title={Predicting Second-Hand Housing Prices in Beijing: A Comparative Study of Machine Learning and Ensemble Models},
booktitle={Proceedings of the 2nd International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI},
year={2025},
pages={574-582},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0014363000004718},
isbn={978-989-758-792-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI
TI - Predicting Second-Hand Housing Prices in Beijing: A Comparative Study of Machine Learning and Ensemble Models
SN - 978-989-758-792-4
AU - He J.
PY - 2025
SP - 574
EP - 582
DO - 10.5220/0014363000004718
PB - SciTePress