{"id":"W4240834168","doi":"10.1002/sam.10016","title":"Discovering and Exploiting Statistical Properties for Query Optimization in Relational Databases: A Survey","year":2009,"lang":"en","type":"article","venue":"Statistical Analysis and Data Mining The ASA Data Science Journal","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Query optimization; Relational database; Data mining; Relational database management system; Cardinality (data modeling); Relational model; Sargable; View; Information retrieval; Key (lock); Exploratory data analysis; Database; Database design; Web search query; Search engine","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":[0.01510498,0.001243985,0.002831351,0.007893068,0.0006763841,0.004725107,0.003522962,0.001133015,0.0009160828],"category_scores_gemma":[0.03886917,0.00122129,0.002275746,0.01070652,0.001963967,0.006773667,0.00191047,0.001439773,0.0006080469],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001273217,"about_ca_system_score_gemma":0.002299574,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00268688,"about_ca_topic_score_gemma":0.001658349,"domain_scores_codex":[0.9850631,0.004492248,0.001897105,0.002019051,0.00624076,0.0002878491],"domain_scores_gemma":[0.942894,0.0414012,0.003871389,0.006059011,0.005329917,0.0004446704],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001773373,0.0003393748,0.03035923,0.002994521,0.0005229906,0.000310164,0.0006135264,0.06085404,0.005550044,0.04931622,0.005040434,0.8439221],"study_design_scores_gemma":[0.00007955518,0.0003201259,0.01219477,0.0005842703,0.0003304623,0.001493433,0.0006876361,0.7886672,0.0170607,0.1385567,0.03980939,0.0002158339],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.02105964,0.03269039,0.9401305,0.001684899,0.0000504741,0.0001404281,0.000517801,0.001565379,0.002160447],"genre_scores_gemma":[0.2500878,0.02828067,0.7178424,0.000369433,0.0004562394,0.0002617213,0.001698093,0.0004292568,0.0005744135],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.01510498,"threshold_uncertainty_score":0.07988369,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1529843466621269,"score_gpt":0.3608533658037779,"score_spread":0.207869019141651,"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."}}