{"id":"W2884303974","doi":"10.1111/1755-0998.12926","title":"Supervised machine learning outperforms taxonomy‐based environmental <scp>DNA</scp> metabarcoding applied to biomonitoring","year":2018,"lang":"en","type":"article","venue":"Molecular Ecology Resources","topic":"Environmental DNA in Biodiversity Studies","field":"Environmental Science","cited_by":165,"is_retracted":false,"has_abstract":true,"ca_institutions":"Cytodiagnostics (Canada)","funders":"Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; Deutsche Forschungsgemeinschaft; Carl-Zeiss-Stiftung","keywords":"Biology; Environmental DNA; Taxonomic rank; Bioindicator; Biological classification; Amplicon; Biodiversity; Taxonomy (biology); Taxon; Computational biology; Ecology; Machine learning; Evolutionary biology; Computer science; Genetics; Polymerase chain reaction","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":["metaepi_narrow","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.000348284,0.0004110635,0.0003751364,0.0001479398,0.0009358189,0.00005503111,0.0006086746,0.0001727082,0.0009811837],"category_scores_gemma":[0.0000691141,0.0004085025,0.0001372723,0.0002576077,0.000877067,0.0001152617,0.001287822,0.0002972251,0.004110517],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004273282,"about_ca_system_score_gemma":0.000002968976,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006839901,"about_ca_topic_score_gemma":0.00004414113,"domain_scores_codex":[0.9974436,0.0001260336,0.0003068079,0.0008579333,0.0004462002,0.0008194397],"domain_scores_gemma":[0.9990695,0.0001298473,0.0001272435,0.0003912059,0.000003203292,0.0002790183],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00001504429,0.00009079502,0.7051771,0.000004404068,0.00005900634,0.00002816885,0.0007912448,0.001050587,0.2915128,0.000001836683,0.000201606,0.001067308],"study_design_scores_gemma":[0.0009789071,0.0005957317,0.590604,0.000008570498,0.00009981755,0.000008706777,0.001656327,0.0009646213,0.3119303,0.000009891875,0.0928005,0.0003426198],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9931813,0.0000832897,0.000412059,0.0001220532,0.0001748209,0.0006228658,0.00002110092,0.0001256119,0.005256845],"genre_scores_gemma":[0.9891275,0.00001173711,0.009233461,0.0009003739,0.00008642329,0.0001278999,0.0000270518,0.00004438714,0.0004411368],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1145731,"threshold_uncertainty_score":0.9999321,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01276858083435116,"score_gpt":0.1961043850239808,"score_spread":0.1833358041896297,"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."}}