Residential real estate price forecasting in Portugal

• Autor(a) principal: Zarrabi, Sepehr
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/148045
• topic: Forecasting; Machine learning; Econometrics; Panel data; Neural networks; Gradient boosting; Domínio/Área Científica::Ciências Sociais::Economia e Gestão

This paper employs three panel data and seven machine learning methods, including linear and nonlinear models, to perform accurate predictions of house prices for fifty-one parishes in six municipalities of Portugal. To construct the predictive models, nine time series economic factors and two non-time series features are applied as explanatory variables. Finally, the neigh boring parish's lagged house prices per square meter data is added as a predictor to increase the forecasting accuracies. The utilized models are Artificial Neural Network, eXtream Gradient Boosting, Linear regression, Lasso and Ridge regression, Bayesian regression, Polynomial regression, Pooled OLS, Panel OLS, and First Difference OLS.