{"id":"W3184478963","doi":"10.1007/s10651-021-00512-5","title":"Markov regression model for analyzing big data to predict trajectories of repeated categorical outcomes: an application to $$\\hbox {PM}_{2.5}$$ air pollution data","year":2021,"lang":"en","type":"article","venue":"Environmental and Ecological Statistics","topic":"Air Quality Monitoring and Forecasting","field":"Environmental Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of New Brunswick; University of Prince Edward Island","funders":"","keywords":"Big data; Categorical variable; Air quality index; Markov chain; Air pollution; Computer science; Statistics; Econometrics; Data mining; Meteorology; Geography; 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":[0.009095277,0.001068476,0.001954211,0.001354882,0.0008629062,0.001176066,0.003009573,0.001311722,0.005916493],"category_scores_gemma":[0.01793466,0.0008033208,0.002310189,0.001947853,0.0006976334,0.001602669,0.001542234,0.003672055,0.001530688],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001584976,"about_ca_system_score_gemma":0.002726302,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03486726,"about_ca_topic_score_gemma":0.03718674,"domain_scores_codex":[0.9978551,0.001270067,0.0001206539,0.0003846153,0.000215534,0.0001539659],"domain_scores_gemma":[0.9826589,0.01480567,0.0006182783,0.0008646922,0.0007298743,0.0003225186],"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":[0.0005108913,0.0002775323,0.01431714,0.00016085,0.0004971109,0.0002073951,0.0001778421,0.8186471,0.0005998127,0.07042056,0.01073872,0.08344498],"study_design_scores_gemma":[0.00001837813,0.00002503695,0.0004357103,0.000007140596,0.00001625887,0.00001390588,0.000009134251,0.9854058,0.00007440807,0.01351607,0.0004671548,0.00001092465],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02334109,0.0004372131,0.970562,0.001049367,0.0001513626,0.000123311,0.002041703,0.001593479,0.000700365],"genre_scores_gemma":[0.4935647,0.001441717,0.4853406,0.0006153185,0.0003677586,0.001323762,0.007908561,0.0004446909,0.008992874],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03486726,"threshold_uncertainty_score":0.06932861,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.102188745144181,"score_gpt":0.3142721415315576,"score_spread":0.2120833963873766,"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."}}