{"id":"W2009423697","doi":"10.1007/s10115-013-0658-2","title":"Email mining: tasks, common techniques, and tools","year":2013,"lang":"en","type":"article","venue":"Knowledge and Information Systems","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":73,"is_retracted":false,"has_abstract":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Categorization; World Wide Web; Data science; Visualization; Information retrieval; Data mining; 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":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.000269684,0.00008651211,0.0001182465,0.0001241055,0.0001370568,0.001125468,0.0001475569,0.00007648922,0.000003222534],"category_scores_gemma":[0.00001774483,0.00007361711,0.00001386274,0.00016643,0.00002031368,0.007821758,0.00009426787,0.0000626214,0.0001517724],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001846667,"about_ca_system_score_gemma":0.00001565341,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001233378,"about_ca_topic_score_gemma":0.0000041162,"domain_scores_codex":[0.9993973,0.00003775317,0.0002678625,0.00009093162,0.00009219969,0.0001139831],"domain_scores_gemma":[0.9994678,0.00005265635,0.0001025156,0.0001927405,0.0001129001,0.00007141756],"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":[0.000002228769,0.00001244349,0.001167536,0.0002013258,0.00001055204,4.747519e-7,0.007039928,0.000001448544,0.0001184494,0.03513667,0.02678557,0.9295233],"study_design_scores_gemma":[0.0003308432,0.0001528561,0.004331955,0.0001355436,0.000004368319,0.0001818838,0.000699776,0.04502887,0.001005423,0.0001732885,0.9476953,0.0002599056],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4495901,0.005807269,0.1189352,0.0008664399,0.003206076,0.002332732,0.00001075316,0.001700756,0.4175507],"genre_scores_gemma":[0.9985225,0.00006858673,0.0008617189,0.00007741179,0.0001014685,0.00007686704,0.000006866403,0.000002917044,0.0002816151],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9292635,"threshold_uncertainty_score":0.9999115,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01539316849572635,"score_gpt":0.2287722311190385,"score_spread":0.2133790626233122,"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."}}