{"id":"W2151849342","doi":"10.1109/hicss.2003.1174243","title":"Support vector machines for text categorization","year":2003,"lang":"en","type":"article","venue":"","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":146,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Categorization; Text categorization; Sorting; Set (abstract data type); Support vector machine; Artificial intelligence; Vocabulary; Feature (linguistics); Natural language processing; Information retrieval; Process (computing); Feature vector","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":[0.00278022,0.001627902,0.002087573,0.003416009,0.0006423211,0.002614758,0.002107186,0.002061337,0.01038901],"category_scores_gemma":[0.01315567,0.0004352759,0.0009464442,0.005999781,0.0006826837,0.00311157,0.001307306,0.002698463,0.01102323],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008936721,"about_ca_system_score_gemma":0.001120002,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002385283,"about_ca_topic_score_gemma":0.001786532,"domain_scores_codex":[0.9962812,0.001365156,0.0003124421,0.0005207862,0.001375726,0.0001446001],"domain_scores_gemma":[0.9953393,0.002940573,0.0003091439,0.0004846941,0.0008327913,0.00009348035],"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":[0.0001579701,0.0001505094,0.001033974,0.001125599,0.0002155738,0.000158455,0.0001391394,0.04124416,0.001693278,0.04085144,0.04819452,0.8650354],"study_design_scores_gemma":[0.0001199484,0.0001562433,0.00228126,0.0005549477,0.00009367596,0.0003843851,0.0002207822,0.5490111,0.003298292,0.2931138,0.1506246,0.0001408404],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005207237,0.03559077,0.9315372,0.00307668,0.001297556,0.0007499054,0.003380095,0.007585241,0.0115754],"genre_scores_gemma":[0.1116541,0.02115308,0.8378341,0.0007252376,0.002204414,0.001767083,0.008144868,0.0004492958,0.016068],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01038901,"threshold_uncertainty_score":0.03475469,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01982714791796406,"score_gpt":0.2626373654355585,"score_spread":0.2428102175175944,"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."}}