{"id":"W1991215305","doi":"10.1145/2245276.2245306","title":"Semi-supervised document clustering with dual supervision through seeding","year":2012,"lang":"en","type":"article","venue":"","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"Natural Sciences and Engineering Research Council of Canada; Mitacs","keywords":"Cluster analysis; Seeding; Computer science; Document clustering; Pairwise comparison; Artificial intelligence; Cluster (spacecraft); Correlation clustering; Data mining; Pattern recognition (psychology); Machine learning; Engineering","routes":{"ca_aff":true,"ca_fund":true,"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":[0.003457849,0.001160797,0.001809795,0.001595767,0.001412791,0.001772115,0.003061383,0.001718989,0.001578895],"category_scores_gemma":[0.01178473,0.0007911887,0.001047495,0.002142811,0.001914533,0.003007611,0.002837427,0.001817392,0.001166431],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001358777,"about_ca_system_score_gemma":0.001920748,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003504453,"about_ca_topic_score_gemma":0.006875278,"domain_scores_codex":[0.9967145,0.00128069,0.0001861916,0.0009539915,0.000706602,0.0001580312],"domain_scores_gemma":[0.9922729,0.002819872,0.000730166,0.002350224,0.001508405,0.0003184589],"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":[0.0007150494,0.0003878546,0.004706862,0.0003677033,0.0002110715,0.0002986198,0.001086648,0.3937536,0.02220428,0.03403985,0.009885876,0.5323427],"study_design_scores_gemma":[0.00003297545,0.00004633662,0.0002730927,0.00001473954,0.00001546997,0.00006325876,0.00004427804,0.9781616,0.004990631,0.0148485,0.001488625,0.00002050322],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01389963,0.0001386195,0.9839801,0.00008588435,0.00001769638,0.0001066604,0.00007661855,0.0009861151,0.0007087943],"genre_scores_gemma":[0.2553123,0.0001113334,0.7409336,0.0001330526,0.00006501276,0.0003239177,0.0008744033,0.0002453346,0.002000998],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003504453,"threshold_uncertainty_score":0.01828706,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03048652227191006,"score_gpt":0.2641961432442345,"score_spread":0.2337096209723244,"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."}}