{"id":"W6920737760","doi":"10.6084/m9.figshare.13364273","title":"Additional file 6 of Systematic evaluation of supervised machine learning for sample origin prediction using metagenomic sequencing data","year":2020,"lang":"en","type":"article","venue":"Figshare","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Agency of Canada","funders":"","keywords":"Geographic coordinate system; Sample (material); Longitude; Regression; Multivariate statistics; Metagenomics; Identification (biology); Data point; Point (geometry)","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":["metaresearch","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.01073058,0.002281188,0.002018882,0.004080344,0.00145233,0.002538942,0.00450749,0.001929755,0.7934479],"category_scores_gemma":[0.09383106,0.001123993,0.002910421,0.005306662,0.0007029551,0.002868893,0.002061755,0.00171494,0.1514778],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001550565,"about_ca_system_score_gemma":0.003772077,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006908499,"about_ca_topic_score_gemma":0.01826819,"domain_scores_codex":[0.9962213,0.001136879,0.000566407,0.0009020176,0.0008924079,0.0002810379],"domain_scores_gemma":[0.8819351,0.1019532,0.002632817,0.005451007,0.006995421,0.001032595],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007926042,0.0001740049,0.004020195,0.009248913,0.0003740755,0.0001420978,0.000171316,0.001880015,0.0004411563,0.0009875739,0.9670877,0.01468016],"study_design_scores_gemma":[0.0188947,0.00116755,0.05551441,0.01049043,0.001439527,0.0009367364,0.0009811545,0.0261341,0.005672979,0.02307392,0.8550592,0.0006353029],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.0004189814,0.00003347115,0.001584352,0.0001125203,0.00006179351,0.0002422289,0.9941911,0.002786162,0.0005694311],"genre_scores_gemma":[0.01418932,0.0001612607,0.03558048,0.0007179307,0.0001617166,0.0060463,0.9299278,0.006198979,0.007016239],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9892694,"threshold_uncertainty_score":0.2946215,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4138278817457102,"score_gpt":0.3264325852622777,"score_spread":0.08739529648343253,"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."}}