{"id":"W2901065680","doi":"10.1007/s10489-018-1333-9","title":"Model selection and application to high-dimensional count data clustering","year":2018,"lang":"en","type":"article","venue":"Applied Intelligence","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":false,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Burstiness; Cluster analysis; Model selection; Count data; Dirichlet distribution; Multinomial distribution; Novelty; Selection (genetic algorithm); Data mining; Algorithm; Artificial intelligence; Pattern recognition (psychology); Statistics; Mathematics; Poisson distribution","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":[0.01003485,0.001296019,0.002637035,0.003294832,0.002370579,0.002951187,0.003899473,0.002524215,0.002877516],"category_scores_gemma":[0.05032739,0.001186703,0.002416597,0.004737219,0.001289356,0.002377929,0.003029936,0.00318955,0.001365167],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001220552,"about_ca_system_score_gemma":0.002155557,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009120094,"about_ca_topic_score_gemma":0.01116936,"domain_scores_codex":[0.9946141,0.003712134,0.0002337437,0.0004519334,0.0008444951,0.0001436668],"domain_scores_gemma":[0.9753212,0.02013175,0.0006682695,0.001442502,0.002123637,0.000312556],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002314383,0.0002077476,0.003659608,0.0003215926,0.0003909528,0.0002100332,0.0005957352,0.6575437,0.001987336,0.09242067,0.004457248,0.237974],"study_design_scores_gemma":[0.0000108966,0.00001006474,0.000196829,0.000009678388,0.000015652,0.00003713741,0.00002346298,0.962209,0.0003324712,0.03656328,0.0005763334,0.00001506982],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004018037,0.0002032947,0.9949648,0.0001642307,0.00002329576,0.00003235495,0.00003774465,0.0002963647,0.000259887],"genre_scores_gemma":[0.1271162,0.000655058,0.8689228,0.0001329981,0.0001150297,0.0004230904,0.0004865573,0.0002961349,0.001852145],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01003485,"threshold_uncertainty_score":0.05306995,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04075433305424717,"score_gpt":0.3080225558592062,"score_spread":0.267268222804959,"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."}}