{"id":"W1982069553","doi":"10.5539/cis.v8n1p151","title":"An Empirical Analysis of Imbalanced Data Classification","year":2015,"lang":"en","type":"article","venue":"Computer and Information Science","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Regina","funders":"","keywords":"Computer science; Naive Bayes classifier; Support vector machine; Artificial intelligence; Machine learning; Classifier (UML); Decision tree; Empirical research; Data mining; Pattern recognition (psychology); Statistics; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"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":[0.01762388,0.001289778,0.001086429,0.004459959,0.0008738746,0.001596847,0.001004888,0.001117918,0.001107212],"category_scores_gemma":[0.0744095,0.0002920097,0.0005539202,0.003758314,0.001826284,0.003150963,0.001444026,0.001318655,0.0003998863],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001268419,"about_ca_system_score_gemma":0.0005214578,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009239122,"about_ca_topic_score_gemma":0.0008300624,"domain_scores_codex":[0.9847749,0.005572078,0.001126895,0.002008299,0.00597805,0.000539833],"domain_scores_gemma":[0.9179787,0.05408049,0.007615243,0.01150966,0.008036886,0.0007790014],"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":[0.00142321,0.001081202,0.3785383,0.001384531,0.0008557225,0.0008595975,0.001067809,0.1727738,0.007309596,0.01572471,0.03112063,0.3878608],"study_design_scores_gemma":[0.00008783732,0.001173027,0.1646145,0.000307418,0.0001433504,0.00227895,0.001064126,0.7730291,0.01467484,0.02413529,0.01837902,0.0001124834],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8020816,0.00751128,0.1739679,0.002282324,0.0005608271,0.0003681291,0.004622645,0.001290839,0.007314536],"genre_scores_gemma":[0.9634605,0.001074085,0.02893737,0.0002284992,0.0002357718,0.0001686992,0.004742992,0.0001251382,0.001026896],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01762388,"threshold_uncertainty_score":0.09320515,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1120192065287858,"score_gpt":0.3778644614599534,"score_spread":0.2658452549311676,"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."}}