{"id":"W159079147","doi":"10.1385/1-59259-802-1:301","title":"Comparison of Methods Based on Diversity and Similarity for Molecule Selection and the Analysis of Drug Discovery Data","year":2004,"lang":"en","type":"article","venue":"Methods in molecular biology","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Mitacs","keywords":"Cluster analysis; Similarity (geometry); Chemical space; Set (abstract data type); Data mining; Computer science; Selection (genetic algorithm); Data set; Consensus clustering; Drug discovery; Artificial intelligence; Bioinformatics; Fuzzy clustering; Biology; CURE data clustering algorithm","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.007356655,0.0001459652,0.0006536844,0.0004175647,0.0001403029,0.0000226688,0.0008019578,0.00008601647,5.887597e-7],"category_scores_gemma":[0.00155282,0.0001141498,0.0001220524,0.001134578,0.0003213745,0.0001323495,0.001711922,0.0001585277,1.465977e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003447312,"about_ca_system_score_gemma":0.00006869384,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003147067,"about_ca_topic_score_gemma":0.00005653446,"domain_scores_codex":[0.9944955,0.004228171,0.0004069762,0.0005843248,0.0001180225,0.0001670736],"domain_scores_gemma":[0.9943736,0.004570112,0.0002680729,0.0006697549,0.00008587177,0.00003265818],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005171485,0.0004094698,0.02632442,0.0001101709,0.0007644462,0.000001398472,0.001702366,0.6053668,0.04141847,0.2744502,0.000003952317,0.0489311],"study_design_scores_gemma":[0.001149635,0.00009568314,0.01001886,0.0000101612,0.0002483251,6.654331e-7,0.00003142166,0.8746535,0.05579866,0.05787234,0.00001332782,0.0001074549],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.07714444,0.00030449,0.9216829,0.0004483464,0.00006948809,0.0002865859,0.00003395233,0.000009926864,0.0000198635],"genre_scores_gemma":[0.2834426,0.000007365444,0.7163701,0.0001456426,0.00000325517,0.000007884588,0.00001858408,0.000003720822,7.630373e-7],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2692866,"threshold_uncertainty_score":0.4654895,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07923237722257018,"score_gpt":0.4689759517498672,"score_spread":0.389743574527297,"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."}}