{"id":"W2983385682","doi":"10.1016/j.envsoft.2020.104926","title":"The proper care and feeding of CAMELS: How limited training data affects streamflow prediction","year":2020,"lang":"en","type":"article","venue":"Environmental Modelling & Software","topic":"Hydrology and Watershed Management Studies","field":"Environmental Science","cited_by":243,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"Canada First Research Excellence Fund","keywords":"Streamflow; Training (meteorology); Computer science; Training set; Tree (set theory); Set (abstract data type); Machine learning; Data set; Sequence (biology); Predictive modelling; Data mining; Artificial intelligence; Mathematics; Meteorology; Geography; Drainage basin; Cartography","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001607409,0.0001754043,0.0001612842,0.00001240836,0.0005084836,0.0000306694,0.000323621,0.00006142705,0.00003164027],"category_scores_gemma":[0.00002994937,0.0001297972,0.00003107204,0.00006334629,0.0005171092,0.0003283388,0.0008985329,0.0001609659,0.0000151174],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005011508,"about_ca_system_score_gemma":0.000002392457,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000149236,"about_ca_topic_score_gemma":0.000005849002,"domain_scores_codex":[0.9987927,0.00005218526,0.0001370908,0.0004792727,0.0002503431,0.0002884138],"domain_scores_gemma":[0.9994612,0.00007777032,0.00008154046,0.0002944162,0.00000117784,0.00008384533],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001982682,0.00009979709,0.4250336,0.0001941452,0.0002905312,0.00002400201,0.05181936,0.4053329,0.004541892,0.00002281824,0.002125959,0.1103168],"study_design_scores_gemma":[0.002178346,0.001305015,0.02819369,0.0001705649,0.0005191162,0.00001228681,0.0344644,0.9005799,0.002921557,0.0002645437,0.02832909,0.001061456],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9630593,0.00106937,0.03368945,0.001172148,0.00009124916,0.0004657911,0.0001264637,0.00009552969,0.0002307078],"genre_scores_gemma":[0.9955777,0.000724654,0.003311467,0.0001424651,0.00004713098,0.00001593552,0.00009464686,0.00002088078,0.00006513957],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4952471,"threshold_uncertainty_score":0.5292976,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04101356804679195,"score_gpt":0.1996113489572896,"score_spread":0.1585977809104976,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}