{"id":"W2148900457","doi":"10.1109/icdm.2008.43","title":"SCS: A New Similarity Measure for Categorical Sequences","year":2008,"lang":"en","type":"article","venue":"","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Categorical variable; Similarity (geometry); Computer science; Matching (statistics); Task (project management); Data mining; Artificial intelligence; Pattern recognition (psychology); Range (aeronautics); Sequence (biology); Measure (data warehouse); Similarity measure; Noise (video); Process (computing); Natural language processing; Machine learning; Mathematics; Statistics; Image (mathematics)","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":[0.002262893,0.0007095902,0.001401883,0.01470947,0.0008942795,0.00183576,0.001632651,0.001256595,0.002567466],"category_scores_gemma":[0.01772569,0.0002044145,0.001089274,0.0143564,0.001424373,0.005240633,0.002117217,0.001013613,0.001090192],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001130363,"about_ca_system_score_gemma":0.001665727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001915745,"about_ca_topic_score_gemma":0.001783452,"domain_scores_codex":[0.9945257,0.0007540247,0.0009692695,0.000722228,0.002833881,0.0001949241],"domain_scores_gemma":[0.9880953,0.005378303,0.001730587,0.001147765,0.003172197,0.0004758859],"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":[0.001065371,0.0003537908,0.05842048,0.00176062,0.0006281713,0.0004452109,0.00130048,0.04058157,0.02534102,0.05424673,0.01583873,0.8000178],"study_design_scores_gemma":[0.0002052884,0.001938642,0.04745434,0.0004489965,0.0003229267,0.002982155,0.00182818,0.6569695,0.02227236,0.204001,0.06122075,0.0003558924],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1349193,0.003861274,0.8459577,0.0004334528,0.0004419319,0.0006233507,0.005632217,0.001837657,0.006293056],"genre_scores_gemma":[0.5070941,0.001141843,0.4786527,0.000279524,0.0004866923,0.0009601855,0.00888536,0.0002074182,0.002292148],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01470947,"threshold_uncertainty_score":0.01196748,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07981175858382283,"score_gpt":0.2826599722391048,"score_spread":0.2028482136552819,"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."}}