{"id":"W167522204","doi":"","title":"Beyond TFIDF weighting for text categorization in the vector space model","year":2005,"lang":"en","type":"article","venue":"","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":232,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université Laval","funders":"","keywords":"Weighting; tf–idf; Vector space model; Computer science; Artificial intelligence; Categorization; Leverage (statistics); Vocabulary; Feature selection; Support vector machine; Text categorization; Word (group theory); Feature vector; Natural language processing; Machine learning; Information retrieval; Pattern recognition (psychology); Mathematics; Term (time)","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":[0.004120719,0.001110363,0.001611814,0.003532721,0.0007365138,0.002281934,0.00143613,0.001757777,0.003778939],"category_scores_gemma":[0.01265271,0.0002978698,0.001156874,0.005622689,0.0009035217,0.006875532,0.0007987497,0.001457312,0.002531298],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001252705,"about_ca_system_score_gemma":0.0009194022,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003041183,"about_ca_topic_score_gemma":0.002228802,"domain_scores_codex":[0.997409,0.001035793,0.0001810076,0.000384453,0.0008324778,0.0001573406],"domain_scores_gemma":[0.9953964,0.002480276,0.0002841828,0.0008226074,0.0009138953,0.0001025363],"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":[0.000216534,0.0001735101,0.001849657,0.0003874978,0.0001277073,0.0001021339,0.000230868,0.06175534,0.008326645,0.09910979,0.008145613,0.8195748],"study_design_scores_gemma":[0.0000298901,0.0001337385,0.001124345,0.00006139974,0.00004522824,0.0002456597,0.00006912193,0.7415646,0.005202009,0.2354152,0.01603577,0.00007302973],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006951607,0.001867158,0.9883196,0.0003060808,0.0001719075,0.00008246818,0.0001237548,0.0004842509,0.001693144],"genre_scores_gemma":[0.2809679,0.003157531,0.702567,0.0003590392,0.000948815,0.0003813535,0.0009544413,0.0003191872,0.01034477],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004120719,"threshold_uncertainty_score":0.02179271,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02088681033409115,"score_gpt":0.2605201725384578,"score_spread":0.2396333622043666,"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."}}