{"id":"W4254656533","doi":"10.1007/978-1-4614-6170-8_100587","title":"Categorical Data Similarity","year":2014,"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); Artificial intelligence; Computer science; Natural language processing; Mathematics; Statistics","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":[0.001072195,0.0006677051,0.001094322,0.004087241,0.0008254184,0.002532405,0.001775974,0.0007787926,0.01468685],"category_scores_gemma":[0.006389774,0.000348522,0.001030116,0.007255675,0.0009104576,0.004192028,0.002614894,0.001489045,0.00741073],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001052772,"about_ca_system_score_gemma":0.001125118,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007163061,"about_ca_topic_score_gemma":0.0007489834,"domain_scores_codex":[0.9977088,0.0002951327,0.0002099958,0.0005773287,0.001131109,0.00007756436],"domain_scores_gemma":[0.9981153,0.0004411345,0.00009898076,0.0005866365,0.0006746044,0.00008328007],"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":[0.00003690836,0.00004999834,0.0006804642,0.0003832528,0.00004373861,0.00005992155,0.0001657953,0.00198763,0.002263279,0.1301405,0.0499288,0.8142597],"study_design_scores_gemma":[0.00001288997,0.0001255304,0.002024628,0.0002572213,0.00006528205,0.001616939,0.0003268613,0.03921966,0.008862564,0.4974432,0.4499709,0.00007432712],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007184263,0.01079167,0.8941882,0.002079386,0.001901339,0.0003268711,0.001956872,0.002593654,0.0789777],"genre_scores_gemma":[0.1353667,0.01125732,0.7536694,0.001314342,0.00182771,0.0005166824,0.009365703,0.0007746091,0.08590742],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01468685,"threshold_uncertainty_score":0.04913241,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08237857441381383,"score_gpt":0.2816371582400747,"score_spread":0.1992585838262609,"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."}}