{"id":"W2568188511","doi":"10.1016/j.jda.2011.10.002","title":"Indexability, concentration, and VC theory","year":2011,"lang":"en","type":"article","venue":"Journal of Discrete Algorithms","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Search engine indexing; Histogram; Measure (data warehouse); Dimension (graph theory); Similarity (geometry); Similarity measure; Statistical analysis","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001336821,0.00009544843,0.0001865277,0.00003544967,0.00005648362,0.00006983573,0.0004179185,0.00005439171,0.00002363994],"category_scores_gemma":[0.00006590241,0.00006697478,0.00007296098,0.0001161769,0.0000876007,0.0006573407,0.00009513786,0.0001906588,0.000001477988],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001438035,"about_ca_system_score_gemma":0.00007049488,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005161085,"about_ca_topic_score_gemma":6.001929e-7,"domain_scores_codex":[0.9989542,0.0002137758,0.0003322,0.0001418961,0.0002051929,0.000152752],"domain_scores_gemma":[0.9991623,0.00008451677,0.0002337179,0.0002175578,0.0001530442,0.0001488203],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00003404365,0.00007375014,0.002141413,0.00001407771,0.00007580574,0.0000622523,0.004958413,7.095967e-7,0.0003977638,0.4958844,0.0003641212,0.4959933],"study_design_scores_gemma":[0.0006634151,0.0004167293,0.02245072,0.00003961962,0.00003218763,0.0003988221,0.00007156155,0.004794446,0.004034993,0.966285,0.000614646,0.0001978369],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005721329,0.0007840325,0.9896167,0.0002730578,0.0003369655,0.00005450453,0.000001200305,0.00001137375,0.003200835],"genre_scores_gemma":[0.3334849,0.00009324859,0.666035,0.00020589,0.00011841,6.74694e-7,1.108387e-7,0.000004778416,0.00005699204],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4957955,"threshold_uncertainty_score":0.2731152,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02544690606395558,"score_gpt":0.2651457390474658,"score_spread":0.2396988329835102,"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."}}