{"id":"W2073825221","doi":"10.1007/s10115-007-0090-6","title":"Robust projected clustering","year":2007,"lang":"en","type":"article","venue":"Knowledge and Information Systems","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":70,"is_retracted":false,"has_abstract":false,"ca_institutions":"Simon Fraser University; University of Alberta","funders":"Yale University","keywords":"Cluster analysis; Categorical variable; Disjoint sets; Linear subspace; Outlier; Data mining; CURE data clustering algorithm; Cluster (spacecraft); Subspace topology; Set (abstract data type); Computer science; Data set; Correlation clustering; Clustering high-dimensional data; Single-linkage clustering; Algorithm; Pattern recognition (psychology); Mathematics; Artificial intelligence; Machine learning; Combinatorics","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":[0.003004806,0.002183257,0.003138893,0.003160511,0.00166732,0.003524109,0.003524566,0.003103825,0.009476632],"category_scores_gemma":[0.01216205,0.001301326,0.003427071,0.002928799,0.001752965,0.003090538,0.004389039,0.002712828,0.007641758],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001307331,"about_ca_system_score_gemma":0.002715041,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004153752,"about_ca_topic_score_gemma":0.00433531,"domain_scores_codex":[0.9941738,0.001791817,0.0002755616,0.001946718,0.001520249,0.0002918363],"domain_scores_gemma":[0.9950842,0.001082262,0.0002815233,0.00187352,0.001496609,0.0001819811],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0005996759,0.0001575263,0.0008816859,0.000494365,0.0005761574,0.0001310083,0.0001821717,0.3006151,0.01028872,0.1031293,0.02921272,0.5537316],"study_design_scores_gemma":[0.00002510893,0.00004919743,0.0004786802,0.00003260446,0.00005059741,0.0001085064,0.00004513581,0.9114035,0.00388268,0.07660711,0.007279193,0.00003754938],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003052884,0.0003623419,0.9921829,0.0001997578,0.00009930063,0.00007035005,0.0003887304,0.001068518,0.00257517],"genre_scores_gemma":[0.1651687,0.0007603678,0.8101547,0.0003957471,0.0003207678,0.0004189034,0.005527291,0.001208324,0.01604522],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009476632,"threshold_uncertainty_score":0.03170252,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1299341584742713,"score_gpt":0.3893189388673293,"score_spread":0.259384780393058,"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."}}