{"id":"W2142113519","doi":"10.1145/1081870.1081934","title":"Scalable discovery of hidden emails from large folders","year":2005,"lang":"en","type":"article","venue":"","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Automatic summarization; Scalability; World Wide Web; Popularity; Thread (computing); Robustness (evolution); Information retrieval; Electronic mail; Database","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":[0.002380307,0.001160747,0.002265693,0.003483378,0.001082677,0.002010379,0.00210355,0.001531013,0.001676409],"category_scores_gemma":[0.01532442,0.001065089,0.001043116,0.002872702,0.0007217698,0.004466605,0.002082359,0.001217784,0.001520068],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001026276,"about_ca_system_score_gemma":0.001193178,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005153853,"about_ca_topic_score_gemma":0.006389831,"domain_scores_codex":[0.9977931,0.0004295231,0.0001617043,0.0004733422,0.0008839673,0.0002583796],"domain_scores_gemma":[0.9900441,0.004584556,0.001121486,0.002185865,0.001680305,0.0003837278],"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":[0.001178124,0.000671939,0.02346801,0.0003670346,0.0002145087,0.0007930313,0.0008960328,0.1661132,0.04434146,0.003615478,0.0156877,0.7426535],"study_design_scores_gemma":[0.00002966889,0.00005162769,0.003619642,0.000008128896,0.00002993249,0.0001558999,0.0001906643,0.9820226,0.00753743,0.005236884,0.001100021,0.00001751306],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.4396858,0.000966815,0.539246,0.001209608,0.0001213489,0.000365128,0.001964576,0.01400676,0.002433966],"genre_scores_gemma":[0.6315702,0.0003599956,0.3569929,0.0001349315,0.0002528891,0.0002154944,0.006936701,0.0005209742,0.003016032],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005153853,"threshold_uncertainty_score":0.01258844,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0111579982347704,"score_gpt":0.2214932292112642,"score_spread":0.2103352309764938,"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."}}