{"id":"W2783287373","doi":"","title":"Quantifying duplication to improve data quality.","year":2017,"lang":"en","type":"article","venue":"Conference of the Centre for Advanced Studies on Collaborative Research","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"McMaster University","funders":"","keywords":"Computer science; Data deduplication; Quality (philosophy); Gene duplication; 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":["metaresearch","sts","open_science"],"consensus_categories":[],"category_scores_codex":[0.01406317,0.0001632055,0.0004738764,0.0001678403,0.002026575,0.0006268257,0.006676063,0.00004185745,0.00001852687],"category_scores_gemma":[0.1268217,0.0001000121,0.00006379611,0.000784312,0.0007972142,0.0007992816,0.005344437,0.0001966661,0.00008098473],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001145677,"about_ca_system_score_gemma":0.0002963987,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003198516,"about_ca_topic_score_gemma":0.002533208,"domain_scores_codex":[0.9949221,0.0008526142,0.0007330593,0.0009991754,0.002043249,0.0004498218],"domain_scores_gemma":[0.9764231,0.005693792,0.0009244687,0.007029389,0.009792272,0.0001369395],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001148527,0.0003024879,0.0009725575,0.0001922824,0.0002643657,7.737308e-7,0.006637512,0.0002118593,0.01653377,0.622658,0.1021703,0.2489077],"study_design_scores_gemma":[0.00196194,0.0006146349,0.0092317,0.0005128576,0.00002851211,5.582374e-8,0.1425168,0.001237424,0.04557958,0.09258925,0.7052678,0.0004594547],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.2709593,0.001706497,0.03277018,0.610738,0.00876677,0.02556598,0.02241349,0.0001259989,0.02695371],"genre_scores_gemma":[0.9879181,0.0002858683,0.002625727,0.0001459885,0.00007133011,0.0001761056,0.0000223788,0.00001157828,0.008742941],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7169588,"threshold_uncertainty_score":0.9992726,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.8003330727568276,"score_gpt":0.6486310263808803,"score_spread":0.1517020463759473,"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."}}