{"id":"W2073994361","doi":"10.1039/c0an00345j","title":"Detecting influential observations by cluster analysis and Monte Carlo cross-validation","year":2010,"lang":"en","type":"article","venue":"The Analyst","topic":"Spectroscopy and Chemometric Analyses","field":"Chemistry","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"National Natural Science Foundation of China","keywords":"Principal component analysis; Monte Carlo method; Partial least squares regression; Computer science; Cross-validation; Component (thermodynamics); Cluster (spacecraft); Series (stratigraphy); Sample (material); Data mining; Statistics; Artificial intelligence; Mathematics; Machine learning; Chemistry; Physics","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":[0.02681319,0.001667811,0.001835986,0.004209911,0.001751047,0.001708196,0.002110457,0.001691791,0.001042826],"category_scores_gemma":[0.04908349,0.0007452684,0.001376088,0.00216529,0.001928559,0.001653098,0.0016729,0.002193952,0.0003225924],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001313585,"about_ca_system_score_gemma":0.001802994,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005884987,"about_ca_topic_score_gemma":0.006206884,"domain_scores_codex":[0.9880131,0.00770054,0.0005215206,0.001629598,0.001775984,0.0003592349],"domain_scores_gemma":[0.9562255,0.03281515,0.001840368,0.004701455,0.004046836,0.0003706668],"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":[0.0004744287,0.0002884207,0.01908587,0.000191299,0.0007411899,0.0002263855,0.0005413175,0.7978737,0.005118024,0.0311086,0.002272115,0.1420787],"study_design_scores_gemma":[0.000008952961,0.00002341547,0.0006549949,0.000008235766,0.00001556041,0.00002329373,0.00001735729,0.992668,0.001278998,0.004982758,0.0003039166,0.00001460361],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03560013,0.0001256017,0.9629031,0.00007691148,0.00002874257,0.0001504803,0.00004314138,0.0005385197,0.0005332581],"genre_scores_gemma":[0.4074532,0.00008248988,0.5909427,0.00007964623,0.00003024282,0.000391154,0.0003528696,0.0001589857,0.0005085524],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02681319,"threshold_uncertainty_score":0.1418034,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01336330654662731,"score_gpt":0.2878486748633483,"score_spread":0.2744853683167209,"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."}}