{"id":"W3007670733","doi":"10.3390/s20051261","title":"Designing a Streaming Algorithm for Outlier Detection in Data Mining—An Incrementa Approach","year":2020,"lang":"en","type":"article","venue":"Sensors","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Data stream mining; Sliding window protocol; Anomaly detection; Streaming data; Outlier; Data mining; Context (archaeology); Streaming algorithm; Concept drift; Scalability; Big data; Algorithm; Database; Artificial intelligence; Window (computing)","routes":{"ca_aff":true,"ca_fund":true,"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":[0.002897945,0.001160302,0.001492466,0.002606634,0.000973938,0.001488515,0.002810517,0.001255329,0.001409026],"category_scores_gemma":[0.00947878,0.0006675877,0.001260282,0.00233835,0.0008788585,0.003071527,0.001860662,0.001573721,0.0008514508],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008658433,"about_ca_system_score_gemma":0.001771947,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002877062,"about_ca_topic_score_gemma":0.002161171,"domain_scores_codex":[0.9977508,0.000425848,0.0003232561,0.0006292008,0.0007214129,0.0001494802],"domain_scores_gemma":[0.9955776,0.001433077,0.0003317689,0.000732943,0.001738463,0.0001862078],"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":[0.0004310177,0.0002289059,0.005813087,0.0003185961,0.0001758703,0.0003521528,0.0003440645,0.2436751,0.02429558,0.03377332,0.005123573,0.6854687],"study_design_scores_gemma":[0.00001907442,0.0000884111,0.0002559156,0.00001229254,0.00001389318,0.000135623,0.00003147078,0.9842433,0.004413623,0.007770864,0.003000862,0.00001467315],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002862038,0.0001056531,0.9960496,0.0001017746,0.00002844507,0.00004938355,0.00002549775,0.0005680273,0.0002095456],"genre_scores_gemma":[0.07136811,0.0002032795,0.9270345,0.0001025122,0.00008361761,0.0001850985,0.0002119905,0.000120207,0.0006907456],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.002897945,"threshold_uncertainty_score":0.01532602,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07941839438265871,"score_gpt":0.2957602859902442,"score_spread":0.2163418916075855,"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."}}