Unveiling Switzerland's and Poland's Residential Markets with Boosted Trees and SHAP
Nicola Stalder,
Michael Mayer,
Martin E. Hoesli and
Steven Bourassa
ERES from European Real Estate Society (ERES)
Abstract:
What factors impact residential rents, and how? We use a large dataset of offered rents and of explanatory variables for Switzerland to provide insights concerning the dynamics of the country's residential real estate landscape. Rents are modeled by boosted trees (LightGBM) and analyzed with typical tools from explainable AI. We then focus on an innovative application of SHAP (SHapley Additive exPlanations) values to assess the combined location effect of all geographic variables. Thanks to their additivity, the SHAP values of all geographic variables can be summed up and visualized as heatmaps providing full transperency on local and regional rent levels. We further highlight the impact of considering interaction constraints to bring additional structure to the results and achieving a high interpretability of the model. The methods used for the Swiss market are then applied to a dataset of rents and prices for the largest cities in Poland, including the city of Gdansk.
Keywords: Automated Valuation Models; Boosted Trees; Residential Real Estate; Shapley Additive Explanations (SHAP) (search for similar items in EconPapers)
JEL-codes: R3 (search for similar items in EconPapers)
Date: 2024-01-01
New Economics Papers: this item is included in nep-ure
References: Add references at CitEc
Citations:
Downloads: (external link)
https://eres.architexturez.net/doc/oai-eres-id-eres2024-169 (text/html)
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:arz:wpaper:eres2024-169
Access Statistics for this paper
More papers in ERES from European Real Estate Society (ERES) Contact information at EDIRC.
Bibliographic data for series maintained by Architexturez Imprints ().