Explainable AI in a Real Estate Context - Exploring the Determinants of Residential Real Estate Values

Krämer, Bastian; Stang, Moritz; Nagl, Cathrine; Schäfers, Wolfgang

doi:10.2139/ssrn.3989721

Cited by 5 publications

(12 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Looking at further extensions, the possibilities seem ample. Other global interpretation techniques can be considered, such as permutation FI, partial dependence plots and accumulated local effect plots that Krämer et al (2021) and Lorenz et al (2022) already applied in their research. Krämer et al (2021) applied it for seven major cities in Germany.…”

Section: Discussionmentioning

confidence: 99%

“…Other global interpretation techniques can be considered, such as permutation FI, partial dependence plots and accumulated local effect plots that Krämer et al (2021) and Lorenz et al (2022) already applied in their research. Krämer et al (2021) applied it for seven major cities in Germany. Lorenz et al (2022) applied it for Frankfurt am Main, which is also a major German city.…”

Section: Discussionmentioning

confidence: 99%

“…Notably, most of the studies compare the performance of only a few ML models. However, the publications of Krämer et al (2021) and Lorenz et al (2022) point in the direction that the preferences should tend toward tree-based ensemble models, such as RFR and XGBR, due to their advantages.…”

Section: Machine Learning Modelsmentioning

confidence: 99%

“…These techniques are relatively new in economic applications and certainly in real estate prediction applications. So far, there are two works that use global IML techniques, namely, studies by Krämer et al (2021) and Lorenz et al (2022), which both take a similar approach by using permutation FI, partial dependence plots and accumulated local effect plots for the interpretation of their prediction models. They both concluded that the IML techniques improved the transparency of their black box ML models.…”

Section: Machine Learning Techniquesmentioning

confidence: 99%

“…In the first global interpretation method considered, the SHAP FI plot, the average absolute SHAP values per feature are calculated across the data and then visualized (Molnar, 2022). SHAP FI plots is an alternative to the permutation feature importance used in Krämer et al (2021) and Lorenz et al (2022). SHAP FI considers in a highly compressed manner, which features have the most impact, globally, in making predictions (Molnar, 2022).…”

Section: Interpretable Machine Learning Techniquesmentioning

confidence: 99%

See 4 more Smart Citations

Predictability of Belgian residential real estate rents using tree-based ML models and IML techniques

Lenaers

Boudt

Moor

2023

IJHMA

View full text Add to dashboard Cite

Purpose The purpose is twofold. First, this study aims to establish that black box tree-based machine learning (ML) models have better predictive performance than a standard linear regression (LR) hedonic model for rent prediction. Second, it shows the added value of analyzing tree-based ML models with interpretable machine learning (IML) techniques. Design/methodology/approach Data on Belgian residential rental properties were collected. Tree-based ML models, random forest regression and eXtreme gradient boosting regression were applied to derive rent prediction models to compare predictive performance with a LR model. Interpretations of the tree-based models regarding important factors in predicting rent were made using SHapley Additive exPlanations (SHAP) feature importance (FI) plots and SHAP summary plots. Findings Results indicate that tree-based models perform better than a LR model for Belgian residential rent prediction. The SHAP FI plots agree that asking price, cadastral income, surface livable, number of bedrooms, number of bathrooms and variables measuring the proximity to points of interest are dominant predictors. The direction of relationships between rent and its factors is determined with SHAP summary plots. In addition to linear relationships, it emerges that nonlinear relationships exist. Originality/value Rent prediction using ML is relatively less studied than house price prediction. In addition, studying prediction models using IML techniques is relatively new in real estate economics. Moreover, to the best of the authors’ knowledge, this study is the first to derive insights of driving determinants of predicted rents from SHAP FI and SHAP summary plots.

show abstract

Section: Discussionmentioning

confidence: 99%