{"id":"W2021739502","doi":"10.5555/1182635.1164157","title":"Multi-column substring matching for database schema translation","year":2006,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":41,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Substring; Computer science; Database schema; Schema (genetic algorithms); Schema matching; Data mining; Algorithm; Artificial intelligence; Set (abstract data type); Database; Theoretical computer science; Database design; Information retrieval; Programming language; Data integration","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":[],"consensus_categories":[],"category_scores_codex":[0.00151959,0.0008185359,0.0007680375,0.002952063,0.0007592178,0.00150506,0.001832508,0.001187009,0.007215354],"category_scores_gemma":[0.006466491,0.0005198514,0.0009449517,0.00469109,0.0007735803,0.00332423,0.001123808,0.001220352,0.004373476],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006191254,"about_ca_system_score_gemma":0.001435084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001834272,"about_ca_topic_score_gemma":0.002445116,"domain_scores_codex":[0.9982501,0.0004339754,0.00022285,0.000460171,0.0005475245,0.00008535894],"domain_scores_gemma":[0.9964643,0.001316571,0.0003062965,0.001219435,0.0006175906,0.00007579802],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003281153,0.0001904079,0.003326322,0.0007475654,0.0001375363,0.0003808277,0.0004643284,0.009953081,0.04828793,0.02598039,0.02111521,0.8890883],"study_design_scores_gemma":[0.0001273631,0.0003086694,0.003526597,0.0002454107,0.0002150606,0.002931122,0.0005161929,0.5329255,0.2507503,0.09080526,0.1174775,0.0001709755],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006872103,0.0005492311,0.9826338,0.0001868588,0.00007934889,0.0001365535,0.001117127,0.007229435,0.001195548],"genre_scores_gemma":[0.03323588,0.0002084861,0.9623159,0.0001155724,0.00003006993,0.0000993361,0.002812877,0.0003127146,0.0008691378],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007215354,"threshold_uncertainty_score":0.02413774,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03110730105148172,"score_gpt":0.2964350817566067,"score_spread":0.2653277807051249,"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."}}