Real Estate Appraisal Performance Improvement by Adapting a Hybrid Model: Geographically Weighted Regression and Extreme Gradient Boosting in Al Bireh, Palestine
Real estate appraisal functions as a decision support system in pivotal financial, economic, and business processes, encompassing buying and selling houses, property tax, bank lending, and insurance companies. Therefore, improving appraisal performance, among other daunting challenges such as interp...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Universitas Indonesia
2025-05-01
|
| Series: | International Journal of Technology |
| Subjects: | |
| Online Access: | https://ijtech.eng.ui.ac.id/article/view/7196 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Real estate appraisal functions as a decision support system in pivotal financial, economic, and business processes, encompassing buying and selling houses, property tax, bank lending, and insurance companies. Therefore, improving appraisal performance, among other daunting challenges such as interpretability, stability, generalizability, data availability, and evaluation metrics, has been a consistent and substantial demand, attracting attention from academia and stakeholders over the years. The original contribution of this research to knowledge is the adaptation of a hybrid model that combines Geographically Weighted Regression (GWR) with Extreme Gradient Boosting (XGBoost)—termed the hybrid GWR-XGBoost model—using Cook’s Distance (Cook’s D), distinguishing this research from other studies aimed at improving performance. The rationale behind this hybrid model is to simultaneously address the nonstationarity and nonlinearity aspects inherent in the real estate industry to improve performance. This aim is achieved by identifying features influencing real estate appraisal, particularly apartments within residential buildings, in the context of Al Bireh city, Palestine, designing the hybrid GWR XGBoost model, and evaluating its performance. The findings indicate that, across five commonly used statistical performance evaluation metrics—namely MSE, RMSE, MAE, MAPE, and R²—the hybrid GWR-XGBoost model significantly outperforms both GWR and XGBoost, achieving an 11% improvement in R². The implications of this research are threefold: first, features influencing residential real estate appraisal are identified, serving as a reference checklist for stakeholders, including buyers, sellers, developers, appraisers, and investors. Second, enhanced appraisal performance aids in making well-informed financial, economic, and business decisions, ultimately affecting various aspects of people's lives. Lastly, within Palestine, this study lays the groundwork for exploring advanced model-based methods as an alternative to traditional manual methods for real estate appraisal. Overall, improving appraisal performance facilitates informed, evidence-based decision-making, maximizes benefits in economic and financial transactions, and impacts both individuals and their environments policymakers. |
|---|---|
| ISSN: | 2086-9614 2087-2100 |