{"id":"W4386195873","doi":"10.1038/s41597-023-02464-w","title":"Digitization of the Australian Parliamentary Debates, 1998–2022","year":2023,"lang":"en","type":"article","venue":"Scientific Data","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Parliament; Digitization; Democracy; Politics; Resource (disambiguation); Metadata; Public administration; Political science; Library science; Data science; Sociology; Computer science; World Wide Web; Law","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.002913886,0.0004318311,0.0004380034,0.02134539,0.002353337,0.002829399,0.001161459,0.0007275153,0.01956034],"category_scores_gemma":[0.01519977,0.0007866721,0.0003610646,0.03767818,0.0007918038,0.001665114,0.002978006,0.001338025,0.005737589],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007411446,"about_ca_system_score_gemma":0.01075133,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.3613218,"about_ca_topic_score_gemma":0.505362,"domain_scores_codex":[0.9967397,0.0003816665,0.0003643459,0.0003429138,0.001901674,0.0002696754],"domain_scores_gemma":[0.988758,0.001766916,0.001812733,0.001761036,0.005128806,0.0007725452],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004488453,0.0001278232,0.0229739,0.003555486,0.00008519444,0.0007448245,0.05058613,0.001237782,0.004007882,0.02236065,0.5980874,0.295784],"study_design_scores_gemma":[0.000008484227,0.00001283059,0.1763573,0.0002679582,0.000017233,0.000130414,0.002018762,0.0002581408,0.0007299423,0.0006183619,0.8195564,0.00002422798],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.1391423,0.006252722,0.003419147,0.005057581,0.001025673,0.0008466785,0.6130989,0.0009899165,0.2301671],"genre_scores_gemma":[0.2584943,0.007036698,0.01787659,0.001769221,0.0005312529,0.002922929,0.4876778,0.0006723448,0.2230187],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.3613218,"threshold_uncertainty_score":0.7184374,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1295125417729573,"score_gpt":0.39690934396953,"score_spread":0.2673968021965727,"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."}}