{"id":"W2796045805","doi":"10.1007/s10479-018-2830-2","title":"A clustering-based feature selection method for automatically generated relational attributes","year":2018,"lang":"en","type":"article","venue":"Annals of Operations Research","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Data mining; Cluster analysis; Feature selection; Selection (genetic algorithm); Set (abstract data type); Table (database); Relational database; Lasso (programming language); Machine learning; 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":[0.001162279,0.0007932436,0.0014258,0.002747703,0.0007799189,0.000985989,0.001747515,0.0007305256,0.002024132],"category_scores_gemma":[0.002892718,0.0003512656,0.001089407,0.002717949,0.0002867701,0.000874784,0.0008121163,0.0009381393,0.0011428],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004259555,"about_ca_system_score_gemma":0.001059702,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003329954,"about_ca_topic_score_gemma":0.005633262,"domain_scores_codex":[0.9989136,0.0001411329,0.00008921063,0.0002829106,0.0004810907,0.00009199956],"domain_scores_gemma":[0.9986204,0.0004828913,0.00007420096,0.0001761907,0.0005923656,0.00005396469],"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":[0.0002915191,0.000249413,0.001597219,0.0001094905,0.0001345793,0.0001442831,0.00007960984,0.01302484,0.03688523,0.002107056,0.008951373,0.9364254],"study_design_scores_gemma":[0.0001026008,0.0001760032,0.004859352,0.00002268357,0.0001263944,0.0005391124,0.00007962867,0.9507919,0.03235021,0.004432278,0.006439225,0.00008057053],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01500264,0.0002233948,0.9813941,0.00007720332,0.00007128804,0.0001153261,0.0003076078,0.002409633,0.0003988703],"genre_scores_gemma":[0.1284958,0.000138904,0.8671352,0.00007441329,0.00008438526,0.00023455,0.001703313,0.0002231304,0.001910264],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003329954,"threshold_uncertainty_score":0.006771445,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2624505622004498,"score_gpt":0.4869737738800862,"score_spread":0.2245232116796365,"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."}}