{"id":"W3088547571","doi":"10.1145/3395027.3419589","title":"Short Text Stream Clustering via Frequent Word Pairs and Reassignment of Outliers to Clusters","year":2020,"lang":"en","type":"article","venue":"","topic":"Data Stream Mining Techniques","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Outlier; Cluster analysis; Word (group theory); Similarity (geometry); Artificial intelligence; Margin (machine learning); Task (project management); Pattern recognition (psychology); Semantic similarity; Natural language processing; Data mining; Image (mathematics); Machine learning; 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":[0.001650807,0.001610378,0.002085806,0.005626878,0.001044554,0.001963294,0.002352181,0.001217267,0.001070778],"category_scores_gemma":[0.00933568,0.000436372,0.001292775,0.005960806,0.0007559323,0.002428895,0.001743118,0.001550437,0.001468366],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007680363,"about_ca_system_score_gemma":0.001183868,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003193917,"about_ca_topic_score_gemma":0.002767358,"domain_scores_codex":[0.9970861,0.0004158215,0.0003189266,0.0009585302,0.001017616,0.0002030257],"domain_scores_gemma":[0.9953432,0.001167527,0.0009641442,0.0006809405,0.001604147,0.0002399153],"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":[0.001430411,0.0004675853,0.01534032,0.0003998432,0.0003595414,0.0008504792,0.0009496091,0.1057674,0.02729881,0.00926088,0.01338816,0.824487],"study_design_scores_gemma":[0.00005432652,0.0001629409,0.003323344,0.00002999832,0.00006800264,0.0004519551,0.0002824718,0.9678593,0.01299768,0.009875303,0.004832065,0.00006260278],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05593851,0.0006798723,0.9387516,0.0002400095,0.000226217,0.0003173545,0.0006818196,0.002581288,0.0005832796],"genre_scores_gemma":[0.3608891,0.000561538,0.6287087,0.000156921,0.0003967495,0.0004167962,0.005316476,0.0004377708,0.003115867],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005626878,"threshold_uncertainty_score":0.008730412,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0269366959402034,"score_gpt":0.2528465641545121,"score_spread":0.2259098682143087,"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."}}