{"id":"W4295878695","doi":"10.1162/qss_a_00211","title":"Assessing the quality of bibliographic data sources for measuring international research collaboration","year":2022,"lang":"en","type":"article","venue":"Quantitative Science Studies","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Data science; Quality (philosophy); Digital library; Data quality; Information retrieval; Conceptual framework; Metric (unit); Engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2491487,0.000831898,0.001851721,0.04592167,0.003725668,0.01731058,0.00282239,0.001969738,0.002298253],"category_scores_gemma":[0.6693656,0.0009177523,0.001944887,0.07601613,0.004433288,0.0130064,0.009909271,0.002139152,0.0004580558],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007927168,"about_ca_system_score_gemma":0.0126116,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005388139,"about_ca_topic_score_gemma":0.006136108,"domain_scores_codex":[0.6222311,0.2279522,0.05713563,0.008072563,0.08214864,0.002459842],"domain_scores_gemma":[0.1528523,0.5805945,0.1012438,0.05624374,0.1061794,0.002886063],"domain_codex":null,"domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000775507,0.0004930596,0.4834048,0.007956148,0.001895537,0.0001661544,0.01532057,0.007322759,0.002525569,0.09677021,0.006529708,0.3768401],"study_design_scores_gemma":[0.0006451096,0.001539616,0.4749429,0.01993044,0.003480162,0.0008036497,0.04624963,0.07936544,0.040561,0.2294533,0.1019308,0.001097833],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4856026,0.01057276,0.4297014,0.01062175,0.0005524071,0.004984461,0.01177402,0.0008814716,0.04530913],"genre_scores_gemma":[0.7792274,0.001321023,0.21366,0.0002834743,0.0001172069,0.002394874,0.002440005,0.0001262185,0.000429796],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9540783,"threshold_uncertainty_score":0.9259333,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.9238412134420332,"score_gpt":0.7257074572657539,"score_spread":0.1981337561762793,"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."}}