{"id":"W2972941956","doi":"10.1109/tkde.2019.2940019","title":"Bayesian Networks for Data Integration in the Absence of Foreign Keys","year":2019,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canada Research Chairs; University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada; Ontario Centres of Excellence","keywords":"Computer science; Relational database; Probabilistic logic; Inference; Data integration; Bayesian network; Data mining; Cardinality (data modeling); Relation (database); Population; Theoretical computer science; Data science; Information retrieval; Machine learning; Artificial intelligence","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02419687,0.001527284,0.002613338,0.00453459,0.00169349,0.005363632,0.00417495,0.002678343,0.002833636],"category_scores_gemma":[0.08909432,0.002722905,0.003411911,0.005167773,0.003512606,0.01280874,0.006533896,0.006499369,0.0008003059],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005082397,"about_ca_system_score_gemma":0.003770118,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.024928,"about_ca_topic_score_gemma":0.02193068,"domain_scores_codex":[0.9794922,0.009880295,0.001248213,0.005177793,0.003609634,0.0005918387],"domain_scores_gemma":[0.9450961,0.04319925,0.003397573,0.005053985,0.00265493,0.0005981139],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001702395,0.00009354457,0.004231642,0.000265082,0.0004890413,0.0003710315,0.0006668621,0.6137443,0.0007190916,0.2960218,0.002605375,0.08062199],"study_design_scores_gemma":[0.00001460766,0.00001150395,0.000360211,0.00003338039,0.00004573506,0.00004761085,0.00003249727,0.7411014,0.0002849596,0.2561222,0.001926455,0.00001944602],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002620208,0.0002481327,0.9954757,0.0005157,0.00001590534,0.00004329275,0.0001681302,0.0002758004,0.0006369895],"genre_scores_gemma":[0.1968804,0.0008120085,0.7967286,0.0008336678,0.0002430873,0.000513001,0.001658144,0.0002598193,0.002071321],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.024928,"threshold_uncertainty_score":0.1279669,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1235244620639133,"score_gpt":0.3687156893233667,"score_spread":0.2451912272594534,"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."}}