{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001106922,0.0006981149,0.00111935,0.004181486,0.000842032,0.00268975,0.001803081,0.0008344551,0.01569534],"category_scores_gemma":[0.00712935,0.0003644137,0.00107449,0.007431558,0.0009358837,0.004519269,0.002705486,0.001627283,0.008477668],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001089886,"about_ca_system_score_gemma":0.001130893,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006478624,"about_ca_topic_score_gemma":0.000669147,"domain_scores_codex":[0.9976235,0.0003006901,0.0002162341,0.0006258559,0.001152589,0.00008110906],"domain_scores_gemma":[0.9979044,0.0004886722,0.0001059217,0.000669729,0.0007361197,0.00009517048],"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":[0.00004001188,0.00005526823,0.0006778807,0.0004291078,0.00004774535,0.0000602975,0.000162377,0.002026756,0.002306205,0.1426861,0.05584911,0.7956593],"study_design_scores_gemma":[0.00001266862,0.0001245214,0.001864277,0.00025236,0.0000644131,0.001498174,0.000288327,0.03524962,0.00818977,0.4932841,0.4591009,0.00007084989],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007074192,0.01190803,0.8862116,0.002387471,0.002414151,0.0003386922,0.002138703,0.002800269,0.08472697],"genre_scores_gemma":[0.1409676,0.01255904,0.7392769,0.001544694,0.002369606,0.0005473117,0.01041046,0.0008864488,0.09143788],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01569534,"threshold_uncertainty_score":0.05250615,"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."}}