XGBoost Model for Predicting Property Prices in the UK Real Estate Market

Nwora, Peter Amaechi, Safieddine, Fadi and Ahad, Md Atiqur Rahman (2025) XGBoost Model for Predicting Property Prices in the UK Real Estate Market. 13th International Conference on Frontiers of Intelligent Computing: Theory and Applications. (In Press)

Abstract

Abstract. This study contributes to the enhancement of the predictive models and mitigating regression problem, using data science machine learning approach. It critically evaluates studies that have been done within the predictive valuation, forecasting models, using the current technological trends in machine learning, through the proposed framework: Sustainable Feature Machine Learning Agile Framework (SFMLAF). SFMLAF suggests that our proposed model based on XGBoost demonstrated better accuracy based on these results; utilizing the following validation metrics: XGBoost RMSE: 0.444, MAPE: 1.94%, MAE: 0.234. The required datasets were sourced online from the UK government property sold dataset, under the Open Government licence v.3.0, focusing on four UK cities with the following data observations: London: 658,337, Peterborough: 44,635, Leeds: 102,984, and Manchester: 154,626. In conclusion, the proposed predictive model aims to deliver practical benefits for real estate professionals, homebuyers and sellers by enhancing the accuracy and reliability of property valuation in the UK market.

Actions (login required)

Edit Item Edit Item