{"id":"W4406458249","doi":"10.1109/bigdata62323.2024.10825314","title":"Common Pitfalls with Data Deduplication Parameters and Metrics","year":2024,"lang":"en","type":"article","venue":"","topic":"Advanced Data Storage Technologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Data deduplication; Computer science; Data mining; Database","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001297116,0.00007386671,0.00007289545,0.0001476597,0.00003620437,0.000250711,0.001260365,0.00003047223,0.000001115583],"category_scores_gemma":[0.0000569339,0.00005182614,0.000004407771,0.0008053476,0.00007097069,0.00141567,0.001133742,0.00009356773,0.00002046761],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001999242,"about_ca_system_score_gemma":0.00001799626,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002790431,"about_ca_topic_score_gemma":0.00001988153,"domain_scores_codex":[0.9992335,0.000008242843,0.00007997454,0.0004391599,0.000125033,0.0001141233],"domain_scores_gemma":[0.9982153,0.0001769715,0.00001903046,0.001547737,0.00001349531,0.00002752528],"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":[0.000001467622,0.00001343588,0.000405283,0.00002233479,0.00001808019,0.0000293247,0.00004529825,0.00003692784,0.0001601429,0.3475914,0.005278851,0.6463975],"study_design_scores_gemma":[0.0002009086,0.0002401623,0.001203615,0.00006071827,0.00002783765,0.000170955,0.0001218338,0.8676829,0.006433444,0.05790507,0.0654396,0.0005129473],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006220891,0.001027745,0.9895214,0.00155347,0.0000490228,0.00009579209,0.00001324199,0.001139489,0.0003789506],"genre_scores_gemma":[0.3770091,0.000102064,0.6226683,0.0001436416,0.000003262269,0.00000597569,0.00001888329,0.000004627729,0.00004415995],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.867646,"threshold_uncertainty_score":0.2417612,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0493737158693585,"score_gpt":0.2980830176485887,"score_spread":0.2487093017792302,"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."}}