Abstract
This study aims to propose an efficient Machine Learning (ML) model, namely Gradient Boosting Regression Trees (GBRT), to predict apartment prices considering the fluctuation of construction material prices and the annual inflation index. For developing the ML model, 480 apartments in Vinh City (Vietnam) were considered. The input parameters employed while training the ML model were the area of the apartments, the number of bedrooms/restrooms, the apartment class, nearby health or education services, investment potential, and parking, whereas the apartment price was the output of the model. The results show that the GBRT model predicts the apartment price accurately with a high value of 0.997 and a small RMSE of 0.26. Additionally, the obtained a20-index is very high, almost 1.0. Finally, a practical graphical user interface was developed to facilitate the prediction of the apartment price in terms of usability.
Publisher
Engineering, Technology & Applied Science Research
Reference20 articles.
1. P. Abelson, R. Joyeux, G. Milunovich, and D. Chung, "Explaining House Prices in Australia: 1970–2003," Economic Record, vol. 81, no. s1, pp. S96–S103, 2005.
2. H. K. Singla and P. Bendigiri, "Factors affecting rentals of residential apartments in Pune, India: an empirical investigation," International Journal of Housing Markets and Analysis, vol. 12, no. 6, pp. 1028–1054, Jan. 2019.
3. M. Kamal and S. A. Pramanik, "Factors Affecting Customers to Buy Apartments in Dhaka City," Daffodil International University Journal of Business and Economics, vol. 9, no. 2, pp. 37–49, Dec. 2015.
4. M. Hilmi, M. Masri, A. H. Nawawi, and I. Sipan, "Review of Building, Locational, Neighbourhood Qualities Affecting House Prices in Malaysia," Procedia - Social and Behavioral Sciences, vol. 234, pp. 452–460, Oct. 2016.
5. H. Selim, "Determinants of house prices in Turkey: Hedonic regression versus artificial neural network," Expert Systems with Applications, vol. 36, no. 2, part 2, pp. 2843–2852, Mar. 2009.