{"id":"W2044838388","doi":"10.1145/2361354.2361393","title":"Personalized document clustering with dual supervision","year":2012,"lang":"en","type":"article","venue":"","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Cluster analysis; Personalization; Matching (statistics); Feature (linguistics); Ask price; Information retrieval; Document clustering; Data mining; Machine learning; Artificial intelligence; World Wide Web","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":[0.004062357,0.0009878724,0.001605849,0.002196652,0.001107084,0.001876479,0.002207428,0.001530109,0.002323243],"category_scores_gemma":[0.01387649,0.0008603885,0.00130225,0.002121191,0.001035601,0.003111605,0.00225979,0.001661019,0.002137377],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001557189,"about_ca_system_score_gemma":0.001725184,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004837904,"about_ca_topic_score_gemma":0.008709313,"domain_scores_codex":[0.9948221,0.001869763,0.0002860412,0.001664142,0.001133018,0.0002248572],"domain_scores_gemma":[0.9899883,0.002967948,0.0008162147,0.004171109,0.001751754,0.0003047587],"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":[0.00097263,0.0005888536,0.007687023,0.000349934,0.00043449,0.0001261725,0.0007639454,0.2326434,0.01508539,0.008161019,0.008180265,0.7250069],"study_design_scores_gemma":[0.00003975502,0.0001014253,0.001433771,0.00001777399,0.0000381395,0.0001222724,0.00008762073,0.975729,0.007563686,0.01260901,0.002223173,0.00003428722],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03134776,0.0002626468,0.9620962,0.0002080644,0.00003405496,0.0001870672,0.0002773973,0.004016758,0.001569942],"genre_scores_gemma":[0.387815,0.0001519814,0.6066604,0.000157434,0.00007564653,0.0002487513,0.001262601,0.0003503118,0.003277955],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004837904,"threshold_uncertainty_score":0.02148402,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01932422136775616,"score_gpt":0.2607877174463604,"score_spread":0.2414634960786042,"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."}}