Please use this identifier to cite or link to this item: https://doi.org/10.1007/s11146-021-09861-1
Title: Boosted Tree Ensembles for Artificial Intelligence Based Automated Valuation Models (AI-AVM)
Authors: Sing, TF 
Yang, JJ 
Yu, SM 
Issue Date: 1-Jan-2021
Publisher: Springer Science and Business Media LLC
Citation: Sing, TF, Yang, JJ, Yu, SM (2021-01-01). Boosted Tree Ensembles for Artificial Intelligence Based Automated Valuation Models (AI-AVM). Journal of Real Estate Finance and Economics. ScholarBank@NUS Repository. https://doi.org/10.1007/s11146-021-09861-1
Abstract: This paper develops an artificial intelligence based automated valuation model (AI-AVM) using the boosting tree ensemble technique to predict housing prices in Singapore. We use more than 300,000 private and public housing transactions in Singapore for the period from 1995 to 2017 in the training of the AI-AVM models. The boosting model is the best predictive model that produce the most robust and accurate predictions for housing prices compared to the decision tree and multiple regression analysis (MRA) models. The boosting AI-AVM models explain 91.33% and 94.28% of the price variances, and keep the mean absolute percentage errors at 8.55% and 5.34% for the public housing market and the private housing market, respectively. When subject the AI-AVM to the out-of-sample forecasting using the 2018 housing sale samples, the prediction errors remain within a narrow range of between 5% and 9%.
Source Title: Journal of Real Estate Finance and Economics
URI: https://scholarbank.nus.edu.sg/handle/10635/224916
ISSN: 08955638
1573045X
DOI: 10.1007/s11146-021-09861-1
Appears in Collections:Staff Publications
Elements

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Sing2021_Article_BoostedTreeEnsemblesForArtific.pdf1.7 MBAdobe PDF

CLOSED

Published

SCOPUSTM   
Citations

4
checked on Sep 23, 2022

Page view(s)

47
checked on Sep 22, 2022

Download(s)

1
checked on Sep 22, 2022

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.