{"id":"W7132862348","doi":"","title":"DTDMatch: dynamically matching streaming documents to DTDs","year":2005,"lang":"","type":"dissertation","venue":"TSpace","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Library and Archives Canada","funders":"","keywords":"Document type definition; XML; Set (abstract data type); Document Structure Description; XML validation; Matching (statistics); Sequence (biology); Identification (biology)","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":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0005346403,0.001085383,0.001066166,0.000459205,0.0007467855,0.0005847359,0.001480292,0.0004105471,0.0003414032],"category_scores_gemma":[0.0001285542,0.001148217,0.0002806072,0.0009614594,0.00005531789,0.001748792,0.0007508326,0.0007923168,0.001329469],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004739614,"about_ca_system_score_gemma":0.0004046111,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002566403,"about_ca_topic_score_gemma":0.003612591,"domain_scores_codex":[0.9941008,0.0001914538,0.001192415,0.001900418,0.001299897,0.00131501],"domain_scores_gemma":[0.9960445,0.0001890644,0.0008062153,0.001913287,0.0003539954,0.0006929658],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004360709,0.0006805339,0.000311893,0.002184645,0.0005465439,0.0004423037,0.2564016,0.01745848,0.03436333,0.2740374,0.002450289,0.4106869],"study_design_scores_gemma":[0.005220221,0.00220651,0.0129088,0.02400683,0.0006414955,0.0004006753,0.2104833,0.06902513,0.0126394,0.003352697,0.6426017,0.01651328],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2100019,0.0005710524,0.7691311,0.000888556,0.003447717,0.001181542,0.00007424604,0.0003297975,0.01437406],"genre_scores_gemma":[0.3147976,0.0003617307,0.5489505,0.000913476,0.001472187,0.0002814568,0.001161476,0.0002828858,0.1317787],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.6401514,"threshold_uncertainty_score":0.9994481,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01477144982025434,"score_gpt":0.3447663779368434,"score_spread":0.3299949281165891,"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."}}