{"id":"W151938044","doi":"10.1137/1.9781611972757.5","title":"Summarizing and Mining Skewed Data Streams","year":2005,"lang":"en","type":"article","venue":"","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":138,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Crohn's and Colitis Foundation of Canada","keywords":"Skew; Skewness; Data stream mining; Computer science; Data mining; Data stream; Space (punctuation); Zipf's law; Point (geometry); Algorithm; Mathematics; Statistics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003105537,0.001117254,0.001743375,0.004433203,0.0007644242,0.002101373,0.001429673,0.0008866058,0.0005868638],"category_scores_gemma":[0.01970888,0.0005009691,0.0007487283,0.004827242,0.0005656896,0.003775259,0.001511983,0.001059103,0.0005536935],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005345318,"about_ca_system_score_gemma":0.0006635269,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009410309,"about_ca_topic_score_gemma":0.001253576,"domain_scores_codex":[0.997635,0.0005164801,0.0003452527,0.0004493259,0.0008919263,0.0001619699],"domain_scores_gemma":[0.9903419,0.004535847,0.001338995,0.00155191,0.001934578,0.0002967235],"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":[0.000855596,0.0002820239,0.04191222,0.0005590905,0.0002814407,0.0007286753,0.0009032156,0.2966869,0.01172053,0.0237361,0.01391959,0.6084146],"study_design_scores_gemma":[0.00002993773,0.00007240969,0.002887233,0.00003074109,0.00003091809,0.0001994187,0.000307864,0.9429041,0.007127322,0.04210357,0.004282346,0.00002407026],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1165912,0.001063474,0.8757625,0.0007775279,0.0001354683,0.0001580818,0.002143906,0.002600747,0.00076709],"genre_scores_gemma":[0.5157912,0.001139075,0.4731952,0.0002152818,0.0004533245,0.0002438265,0.007703266,0.000206904,0.00105193],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004433203,"threshold_uncertainty_score":0.01642382,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04579868955532916,"score_gpt":0.2686245177712225,"score_spread":0.2228258282158933,"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."}}