{"id":"W4243491101","doi":"10.1007/978-1-4939-7131-2_100085","title":"Categorical Data Similarity","year":2018,"lang":"en","type":"book-chapter","venue":"","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Categorical variable; Similarity (geometry); Computer science; Natural language processing; Artificial intelligence; Machine learning","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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0001968598,0.0001928387,0.0001899153,0.00005356683,0.0001095937,0.0002191824,0.004641826,0.0001828382,0.0007333532],"category_scores_gemma":[0.00001420413,0.0001663848,0.00003691517,0.00003964346,0.00008600536,0.0003605098,0.003271188,0.0002322001,0.003836871],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000235481,"about_ca_system_score_gemma":0.0001127952,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002211073,"about_ca_topic_score_gemma":0.0000205732,"domain_scores_codex":[0.9984806,0.000004441155,0.0002099994,0.000871593,0.0002583579,0.0001749932],"domain_scores_gemma":[0.9954563,0.00005186394,0.00008729665,0.00421317,0.00007796684,0.0001134485],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[1.087426e-7,0.00000686143,1.310067e-7,0.000001980291,0.000009468018,0.000004307054,0.000008861968,2.121806e-8,2.581845e-7,0.5856633,0.38148,0.03282475],"study_design_scores_gemma":[0.0000328336,0.00001720883,0.000003143836,0.000005785676,0.00001036991,0.00001426374,5.909238e-7,0.02700828,0.000002779032,0.09617568,0.8765288,0.0002002847],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[4.476266e-8,0.00004104152,0.4788587,0.000693253,0.0001566831,0.00006385595,0.0001632643,0.0001661493,0.519857],"genre_scores_gemma":[0.00001052709,0.00005455405,0.245776,0.0005066388,0.000449886,0.00000402962,0.001022044,0.00001815902,0.7521582],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.4950488,"threshold_uncertainty_score":0.9969388,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.105108445229361,"score_gpt":0.2989304517172668,"score_spread":0.1938220064879058,"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."}}