{"id":"W2083200098","doi":"10.6017/ital.v33i2.5341","title":"The Importance of Identifying and Accommodating E-Resource Usage Data for the Presence of Outliers. The Negative Impacts of Inaccurate E-Journal Usage Data.","year":2014,"lang":"en","type":"article","venue":"Information Technology and Libraries","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Laurentian University","funders":"","keywords":"Outlier; Upload; Computer science; Identification (biology); Sample (material); Usage data; Resource (disambiguation); Data mining; Statistics; Data science; Information retrieval; World Wide Web; Mathematics; Artificial intelligence","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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.04020494,0.0004909662,0.0006475149,0.003325712,0.001503508,0.004875871,0.0009667649,0.001248379,0.001310322],"category_scores_gemma":[0.2542506,0.0003715974,0.000722846,0.004245653,0.002224325,0.00461824,0.001912143,0.001780494,0.0008754639],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001550309,"about_ca_system_score_gemma":0.004403448,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006007868,"about_ca_topic_score_gemma":0.007950366,"domain_scores_codex":[0.9371967,0.02984448,0.004801289,0.002962765,0.02427585,0.0009188588],"domain_scores_gemma":[0.6119809,0.265385,0.05739404,0.0215077,0.04133055,0.00240177],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0004597147,0.0003273095,0.6445698,0.0008906881,0.0004232401,0.000485667,0.005602955,0.008953154,0.005732096,0.007236552,0.01072485,0.314594],"study_design_scores_gemma":[0.00003945339,0.0009953694,0.8766519,0.0009067048,0.000258284,0.001357309,0.01423106,0.03320977,0.01175576,0.02748568,0.03276115,0.0003475471],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7562932,0.00333163,0.1856855,0.02205297,0.0008339173,0.0009886622,0.003407102,0.00103967,0.02636734],"genre_scores_gemma":[0.9517713,0.000488651,0.04430664,0.001117976,0.0002344453,0.000180393,0.0008021205,0.0001146267,0.0009838302],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9951241,"threshold_uncertainty_score":0.2126266,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1325676505123607,"score_gpt":0.3760560784491905,"score_spread":0.2434884279368298,"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."}}