{"id":"W2024888043","doi":"10.1145/2340416.2340421","title":"A unified framework for document clustering with dual supervision","year":2012,"lang":"en","type":"article","venue":"ACM SIGAPP Applied Computing Review","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Cluster analysis; Computer science; Seeding; Document clustering; Correlation clustering; Conceptual clustering; Brown clustering; Data mining; Single-linkage clustering; Artificial intelligence; Clustering high-dimensional data; Pairwise comparison; Constrained clustering; CURE data clustering algorithm; Pattern recognition (psychology); Physics","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":[0.005230008,0.001423587,0.002391804,0.003608054,0.001742528,0.003793706,0.004467894,0.001954849,0.00253956],"category_scores_gemma":[0.008247283,0.0009445209,0.002267514,0.004567406,0.001739546,0.003826719,0.004554408,0.003333246,0.002223231],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002236489,"about_ca_system_score_gemma":0.003867941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006160885,"about_ca_topic_score_gemma":0.01149248,"domain_scores_codex":[0.9951119,0.001780978,0.0002864967,0.001158618,0.001419043,0.000242978],"domain_scores_gemma":[0.9964373,0.0008663432,0.000272207,0.0009640155,0.001227884,0.0002322174],"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":[0.0002032945,0.0002570962,0.001498637,0.000520532,0.000369456,0.000240992,0.0007280104,0.2513456,0.007383936,0.2487435,0.02124657,0.4674625],"study_design_scores_gemma":[0.00002979058,0.00004612638,0.0002660282,0.00003310732,0.00004003576,0.0001125139,0.00006183757,0.8983923,0.001463546,0.0879576,0.01155945,0.00003759728],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0004627901,0.0001394416,0.9984835,0.00007811337,0.00001702238,0.00004281591,0.00006441481,0.0003349191,0.0003770163],"genre_scores_gemma":[0.03341696,0.0003171859,0.9629169,0.0001156838,0.0001202025,0.000304223,0.0007756407,0.0002066016,0.00182671],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006160885,"threshold_uncertainty_score":0.02765924,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03878235484616066,"score_gpt":0.3065368105442365,"score_spread":0.2677544556980758,"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."}}