{"id":"W1546703457","doi":"10.1137/1.9781611972733.6","title":"Hierarchical Document Clustering Using Frequent Itemsets","year":2003,"lang":"en","type":"article","venue":"","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":469,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Cluster analysis; Computer science; Document clustering; Information retrieval; Hierarchical clustering; Brown clustering; Intuition; Vocabulary; Artificial intelligence; Library science; Fuzzy clustering; Canopy clustering algorithm; Psychology","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":[0.002736442,0.001462162,0.002356941,0.01299791,0.001591599,0.002668737,0.001951745,0.001275058,0.001738829],"category_scores_gemma":[0.01408443,0.0007416634,0.001795711,0.01063476,0.0005776742,0.002631754,0.001308696,0.001131503,0.002090389],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001029582,"about_ca_system_score_gemma":0.001428309,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004226096,"about_ca_topic_score_gemma":0.004023103,"domain_scores_codex":[0.9958973,0.001197961,0.000452586,0.0009429297,0.001291705,0.0002175849],"domain_scores_gemma":[0.9937807,0.002857612,0.0007371727,0.0008447085,0.001635085,0.0001446466],"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":[0.0003007356,0.0002485574,0.00747498,0.000778018,0.0005266748,0.000515512,0.0005063927,0.1441382,0.00565214,0.01683213,0.01441965,0.8086071],"study_design_scores_gemma":[0.00005526491,0.0001098917,0.002207426,0.0001126937,0.0001512354,0.0006399642,0.000269226,0.9228911,0.004730357,0.05813573,0.01061628,0.00008077128],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01236632,0.001065179,0.9816865,0.0001966724,0.00008757701,0.0003779845,0.001035102,0.001976786,0.001207935],"genre_scores_gemma":[0.1103045,0.0007689231,0.8834099,0.0000669574,0.0001263706,0.0004533988,0.003752519,0.000123304,0.0009941801],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01299791,"threshold_uncertainty_score":0.01447189,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03053123436791092,"score_gpt":0.2888526014938362,"score_spread":0.2583213671259252,"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."}}