1{2 "record_id": "d8361358e99098ffa7a8408974e7bea4a38e7010",3 "eligible": true,4 "short_ref": "Liu, Chen, Orford, Tian & Zou (2024)",5 "country": "United Kingdom",6 "city": "Cardiff",7 "property_type": "mixed (detached, semi-detached, terraced, flat/maisonette)",8 "price_type": "transaction price (log)",9 "n_obs": 16297,10 "sample_period": "2001-2007",11 "data_source": "HM Land Registry (with floor area), OS street network",12 "methods_compared": [13 "XGBoost",14 "random forest",15 "gradient boosting decision tree"16 ],17 "best_method": "XGBoost",18 "metric": "R2",19 "best_value": 0.76,20 "hedonic_value": null,21 "test_set": "holdout",22 "features_ai": "none",23 "incremental_ai_gain": null,24 "incremental_ai_metric": "",25 "marginal_effects": "SHAP",26 "notes": "Four XGBoost models (closeness/betweenness at 600 m, 3 km, 6 km, 10 km radii); Table 2 reports R2 0.76/0.76/0.76/0.74, MAE 0.18-0.19 and RMSE 0.28-0.30 in log-price units on a 10% random holdout (5-fold CV for tuning). RF and GBDT were said to perform worse but their metrics are not printed; no OLS benchmark. SHAP dependence plots give nonlinear marginal effects of accessibility.",27 "confidence": "high"28}