{"id":"W6958421959","doi":"10.6084/m9.figshare.13364258","title":"Additional file 1 of Systematic evaluation of supervised machine learning for sample origin prediction using metagenomic sequencing data","year":2020,"lang":"en","type":"article","venue":"Figshare","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Agency of Canada","funders":"","keywords":"Metagenomics; Sample (material); Supervised learning; Training set; DNA sequencing; Support vector machine","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":[0.006413567,0.001792425,0.001404241,0.00360475,0.001301867,0.002018474,0.003144363,0.00136964,0.7320232],"category_scores_gemma":[0.06170592,0.0007978232,0.00182812,0.005462025,0.0005876536,0.002132206,0.001751068,0.001444658,0.1277222],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001428291,"about_ca_system_score_gemma":0.003093826,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006860298,"about_ca_topic_score_gemma":0.01732715,"domain_scores_codex":[0.9978515,0.0006021849,0.0003199431,0.0006092011,0.0004577382,0.0001593941],"domain_scores_gemma":[0.9304999,0.05971721,0.001587695,0.003416627,0.004073613,0.0007050158],"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":[0.0007025512,0.0001305766,0.003378442,0.005981449,0.0001955956,0.00009163074,0.00009545095,0.001384879,0.0003205244,0.0008019982,0.9751309,0.0117861],"study_design_scores_gemma":[0.01637377,0.001043157,0.0448237,0.006723734,0.001334628,0.0009190618,0.0009649732,0.01978545,0.005100976,0.02268908,0.8798422,0.0003992228],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"other","genre_scores_codex":[0.0002950781,0.00002338079,0.0008134866,0.00006087991,0.00002450775,0.0001122207,0.9971645,0.001123814,0.0003821155],"genre_scores_gemma":[0.007762991,0.0001055434,0.0150481,0.000320557,0.00006589197,0.002485112,0.9692932,0.001781599,0.003136927],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.9935864,"threshold_uncertainty_score":0.3822363,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2537929523028036,"score_gpt":0.3234077562884753,"score_spread":0.06961480398567177,"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."}}