{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0486078,0.002406122,0.002033531,0.005260498,0.002065884,0.00721576,0.004277971,0.002118059,0.0008779128],"category_scores_gemma":[0.2401421,0.001522283,0.0009359389,0.005567361,0.002674501,0.0104089,0.004570299,0.003849669,0.0008110196],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003195339,"about_ca_system_score_gemma":0.002561625,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001775703,"about_ca_topic_score_gemma":0.002099692,"domain_scores_codex":[0.9157356,0.03075641,0.01931669,0.009824521,0.02267887,0.00168783],"domain_scores_gemma":[0.7058358,0.156154,0.01617303,0.08428396,0.03590748,0.001645705],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00212594,0.0007906043,0.08677908,0.004868246,0.0008350557,0.0009545207,0.005767938,0.0837492,0.05166267,0.04341727,0.03216275,0.6868868],"study_design_scores_gemma":[0.0002740513,0.002141064,0.05360003,0.003151396,0.0005026148,0.005391097,0.005296142,0.3452192,0.3100037,0.16048,0.1125447,0.001396135],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2538318,0.02298196,0.6574472,0.007011461,0.001585408,0.002032978,0.007934971,0.03370325,0.01347107],"genre_scores_gemma":[0.6559236,0.002114393,0.3300824,0.000974471,0.000204069,0.001215869,0.004819438,0.003389043,0.00127669],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0486078,"threshold_uncertainty_score":0.2570657,"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."}}