UQO Working Paper No. 3 — Hedonic housing price models for the US: parametric, quantile, and machine-learning approaches.
1scheme,quantity,point,ci_low,ci_high,boot_se,n_boot2observation bootstrap,R2 OLS,0.6301326374996239,0.6265586217094294,0.633456311233228,0.00172027645106486,10003observation bootstrap,R2 LightGBM,0.8087901494725322,0.80673912293203,0.8107990218966363,0.0010548905026136022,10004observation bootstrap,R2 XGBoost,0.8330320388687278,0.8311074764516817,0.8349226728917114,0.0009794841501841805,10005observation bootstrap,R2 XGBoost - R2 OLS,0.2028994013691039,0.20038104654110833,0.20549984922859993,0.0012991796274595929,10006observation bootstrap,R2 XGBoost - R2 LightGBM,0.024241889396195537,0.023585429301740974,0.02489546659101382,0.00032866747811601395,10007observation bootstrap,R2 LightGBM - R2 OLS,0.17865751197290836,0.17638093069999292,0.18120231701419395,0.0012190168776128583,10008ZIP3-cluster bootstrap,R2 OLS,0.6301326374996239,0.6161519978320202,0.6418738550802072,0.006725571005715354,10009ZIP3-cluster bootstrap,R2 LightGBM,0.8087901494725322,0.7972247281694308,0.8180876100363337,0.005476004639813567,100010ZIP3-cluster bootstrap,R2 XGBoost,0.8330320388687278,0.8224463430072542,0.8415248194814269,0.004974356317099438,100011ZIP3-cluster bootstrap,R2 XGBoost - R2 OLS,0.2028994013691039,0.19338240200397602,0.21300161344296217,0.005123563966821378,100012ZIP3-cluster bootstrap,R2 XGBoost - R2 LightGBM,0.024241889396195537,0.0226254465074817,0.02582732716767391,0.00083192796036477,100013ZIP3-cluster bootstrap,R2 LightGBM - R2 OLS,0.17865751197290836,0.1694962361323362,0.18795418763561786,0.004980200653723421,100014paired t on squared-error differences,OLS minus XGBoost loss,0.13230822436471384,,,0.0009393258290057469,100015paired t on squared-error differences,LightGBM minus XGBoost loss,0.015807840336707984,,,0.00021800639643583365,100016