{"id":"W2058737869","doi":"10.1145/1247715.1247717","title":"Online supervised spam filter evaluation","year":2007,"lang":"en","type":"article","venue":"ACM Transactions on Information Systems","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":117,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Filter (signal processing); Set (abstract data type); Software; Receiver operating characteristic; Confidence interval; Information retrieval; Data mining; Machine learning; Artificial intelligence; Statistics; Mathematics; Operating system","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":[0.009281714,0.001971453,0.002302395,0.002655167,0.001147934,0.001277218,0.002179696,0.003620149,0.003743003],"category_scores_gemma":[0.01641863,0.0003939033,0.0008110997,0.001041538,0.0007611739,0.002085825,0.001292199,0.001240407,0.002505061],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001615393,"about_ca_system_score_gemma":0.001606448,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004482449,"about_ca_topic_score_gemma":0.00614198,"domain_scores_codex":[0.9936571,0.002133443,0.0005261849,0.001220067,0.001855235,0.0006079791],"domain_scores_gemma":[0.9823269,0.007634443,0.001059925,0.002541024,0.005562949,0.0008747604],"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":[0.005432546,0.004897409,0.04432132,0.0006721513,0.001080869,0.0004819783,0.0003759146,0.210568,0.01927669,0.001574801,0.03255678,0.6787615],"study_design_scores_gemma":[0.0002224845,0.001148123,0.01166896,0.00002600667,0.0001271423,0.0003076087,0.0001263018,0.9514154,0.02989076,0.0009477662,0.004062964,0.0000564743],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8385979,0.00159139,0.1340302,0.0004104247,0.0004297313,0.0008312212,0.002160532,0.01426054,0.007687961],"genre_scores_gemma":[0.9022014,0.0001824381,0.08380874,0.0002115775,0.0001432042,0.0003589398,0.005601462,0.0002696189,0.00722253],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009281714,"threshold_uncertainty_score":0.04908699,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03922481280749248,"score_gpt":0.2785531126448064,"score_spread":0.2393282998373139,"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."}}