1{"record_id": "855f8c9e0508b319a7022588b7485215b722768f", "eligible": true, "short_ref": "Tymchuk, Pylypenko & Kicha (2023)", "country": "Ukraine", "city": "Kyiv", "property_type": "apartments in new-build residential complexes", "price_type": "asking price per m2 (developer listings, log)", "n_obs": 169,2 "sample_period": "projects completed 2022-2025 (data year not stated, c. 2023)", "data_source": "lun.ua listing platform (scraped)", "methods_compared": ["OLS hedonic with quantitative regressors (log area, rooms, crime rate, minutes to centre, year)", "OLS with neural-network text-derived housing class (predictedClass) + log area"],3 "best_method": "OLS with NN text-derived housing class", "metric": "R2", "best_value": 0.889, "hedonic_value": 0.884,4 "test_set": "none", "features_ai": "listing text embeddings",5 "incremental_ai_gain": 0.005, "incremental_ai_metric": "R2 (in-sample, 0.884 -> 0.889; model (3) replaces most quantitative regressors by the text-derived class rather than adding to them)", "marginal_effects": "coefficients", "notes": "Ukrainian-language paper with English abstract. A small feed-forward network with nnlm-en-dim50 word embeddings classifies the complex description (translated to English) into 4 housing classes; the predicted class enters a log-linear regression. Model (2) conventional OLS: R2 0.884 (adj 0.880); model (3) with predictedClass + logarea: 'explains about 88.9% of variance' (R2 ~0.889). Both are in-sample fits on 169 records; no holdout. Not a pure ML-vs-hedonic prediction comparison.", "confidence": "medium"}6