{"id":"W2278925444","doi":"10.7939/r3b23s","title":"A Synthetic Data Generator for Clustering and Outlier Analysis","year":2006,"lang":"en","type":"article","venue":"","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":68,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Cluster analysis; Outlier; Data mining; Anomaly detection; Generator (circuit theory); Pattern recognition (psychology); Artificial intelligence","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":[0.003558412,0.0009123929,0.0005614373,0.002071328,0.0004489267,0.000899863,0.001596769,0.0006058586,0.00431211],"category_scores_gemma":[0.01599018,0.0003120587,0.000589792,0.002021965,0.0005068122,0.0008092503,0.00116121,0.0009118824,0.001146454],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005578003,"about_ca_system_score_gemma":0.0007588189,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009468758,"about_ca_topic_score_gemma":0.0008389286,"domain_scores_codex":[0.9978887,0.0007306577,0.0002038824,0.0003032039,0.000793891,0.0000796371],"domain_scores_gemma":[0.9867855,0.006023906,0.0005663616,0.003123677,0.003155261,0.0003452406],"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":[0.001704991,0.001267217,0.02843755,0.001061161,0.0002825223,0.001506862,0.001284841,0.3647825,0.04536338,0.02841318,0.07271301,0.4531827],"study_design_scores_gemma":[0.0001520172,0.0003233228,0.003349463,0.00004072132,0.00003093906,0.0003876761,0.0001318221,0.9263992,0.04078075,0.007299682,0.02103397,0.00007050018],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.07631312,0.00008515522,0.8795829,0.0003154276,0.000252322,0.001002079,0.00856142,0.03076792,0.003119565],"genre_scores_gemma":[0.2859507,0.0001021838,0.6886281,0.00009692132,0.00004292441,0.001601043,0.01950089,0.002502606,0.001574734],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00431211,"threshold_uncertainty_score":0.01881891,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02587532111683141,"score_gpt":0.2705733718957009,"score_spread":0.2446980507788695,"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."}}