{"id":"W4401667657","doi":"10.1007/s10742-024-00335-4","title":"Deduplication methods for literature citations in systematic evidence reviews: practical insights to guide decision-making","year":2024,"lang":"en","type":"article","venue":"Health Services and Outcomes Research Methodology","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Manitoba","funders":"","keywords":"Data deduplication; Computer science; Context (archaeology); Systematic review; Software; Data science; Management science; Data mining; Database; MEDLINE; Engineering","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","scholarly_communication"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.5746341,0.0003389647,0.005186418,0.002633406,0.0004124494,0.001944839,0.001413157,0.0002160315,0.0002282167],"category_scores_gemma":[0.3895076,0.0001614947,0.0007645596,0.006127535,0.000060884,0.0007054969,0.0004044663,0.0006013119,0.0004621519],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002143986,"about_ca_system_score_gemma":0.0004234812,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009436159,"about_ca_topic_score_gemma":0.0009528278,"domain_scores_codex":[0.7580019,0.2170169,0.01683087,0.002852091,0.004141216,0.001156973],"domain_scores_gemma":[0.4113556,0.5789021,0.002231395,0.003779,0.00310485,0.0006270467],"domain_codex":null,"domain_gemma":"methods","domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001122539,0.00009383206,0.003719581,0.1715326,0.0002739391,0.00002691273,0.0531136,0.00008370357,0.0001569251,0.1014112,0.02277734,0.6466981],"study_design_scores_gemma":[0.000277674,0.0004964109,0.01655077,0.08368099,0.0002185129,0.0001425576,0.01378683,0.1072632,0.000006310028,0.3271273,0.4499357,0.0005137764],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001640835,0.1786929,0.7859985,0.02671077,0.0004751969,0.006365437,0.000007948747,0.00001238883,0.00009599533],"genre_scores_gemma":[0.005891031,0.005204247,0.9801933,0.00573009,0.00009535506,0.00191468,0.000005497917,0.00002296136,0.0009428876],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.6461844,"threshold_uncertainty_score":0.9990912,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.9338542154844445,"score_gpt":0.7958391692124195,"score_spread":0.138015046272025,"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."}}