{"id":"W2044679407","doi":"10.1114/1.1289470","title":"Automatic Classification of Protein Sequences into Structure/Function Groups via Parallel Cascade Identification: A Feasibility Study","year":2000,"lang":"en","type":"article","venue":"Annals of Biomedical Engineering","topic":"Machine Learning in Bioinformatics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":24,"is_retracted":false,"has_abstract":false,"ca_institutions":"Queen's University","funders":"","keywords":"Cascade; Artificial intelligence; Weighting; Binary classification; Computer science; Hidden Markov model; Pattern recognition (psychology); Identification (biology); Set (abstract data type); Protein sequencing; Test set; Binary number; Algorithm; Nonlinear system; Machine learning; Peptide sequence; Mathematics; Support vector machine; Biology; Engineering; Genetics","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":[],"consensus_categories":[],"category_scores_codex":[0.0009268625,0.0004216234,0.000450993,0.0005407456,0.0005269251,0.0004818022,0.0006571026,0.0004480522,0.002643895],"category_scores_gemma":[0.00209896,0.0002486833,0.0003406727,0.0004835021,0.0003180432,0.001114187,0.000427112,0.0003505633,0.0005903522],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003033549,"about_ca_system_score_gemma":0.000692639,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002331233,"about_ca_topic_score_gemma":0.002290489,"domain_scores_codex":[0.9996518,0.00007294361,0.00002117555,0.00009248273,0.0001249487,0.00003650794],"domain_scores_gemma":[0.9985266,0.0007242312,0.00006817025,0.0001800643,0.0004284183,0.00007246001],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.004573856,0.00135213,0.01483675,0.0002689978,0.000100972,0.0005278952,0.0001988982,0.03708694,0.3869621,0.00429684,0.001596255,0.5481983],"study_design_scores_gemma":[0.0002195762,0.0008682337,0.006161886,0.000004624079,0.00006956233,0.0005794062,0.00007895913,0.9089425,0.07944684,0.002801012,0.0008061,0.00002141918],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7259208,0.000225307,0.269255,0.0001967929,0.00003018403,0.0002496467,0.0001626724,0.001042328,0.002917216],"genre_scores_gemma":[0.8024974,0.0001270491,0.1953906,0.00003675975,0.00002020745,0.00007418023,0.0002819086,0.00003942125,0.001532594],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.002643895,"threshold_uncertainty_score":0.008844674,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01709792020223794,"score_gpt":0.2927448131064455,"score_spread":0.2756468929042075,"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."}}