{"id":"W1606555179","doi":"10.46430/phen0017","title":"Getting Started with Topic Modeling and MALLET","year":2012,"lang":"en","type":"article","venue":"The Programming Historian","topic":"Geographic Information Systems Studies","field":"Social Sciences","cited_by":117,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo; Carleton University","funders":"","keywords":"Mallet; Computer science; Natural (archaeology); Work (physics); Engineering; History; Mechanical engineering","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":[0.004595008,0.001094921,0.0008827123,0.001934618,0.0009362462,0.004995779,0.001897868,0.001680784,0.04723737],"category_scores_gemma":[0.02074196,0.001058922,0.001783261,0.00274693,0.0009874268,0.01484122,0.003485529,0.007032027,0.03551667],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009519096,"about_ca_system_score_gemma":0.001088159,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002807618,"about_ca_topic_score_gemma":0.005711166,"domain_scores_codex":[0.9974457,0.00105032,0.0001552148,0.0005402573,0.0006775905,0.0001309635],"domain_scores_gemma":[0.9932843,0.00379764,0.0001857128,0.0012352,0.001133288,0.0003638636],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002197477,0.0001081407,0.001590906,0.0005338645,0.0001522077,0.0001914925,0.001563625,0.004122387,0.002011182,0.06451449,0.3202417,0.6047503],"study_design_scores_gemma":[0.0000582798,0.00006009632,0.0009382098,0.0003876424,0.00006231086,0.000406438,0.0005903305,0.02690851,0.004863304,0.1708941,0.7947126,0.0001181427],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001899507,0.002340651,0.9352391,0.01340898,0.001667962,0.0000877482,0.00142179,0.02252227,0.02141206],"genre_scores_gemma":[0.03161709,0.003476764,0.8887871,0.005794323,0.0019683,0.0003115048,0.004653365,0.01193301,0.05145855],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.04723737,"threshold_uncertainty_score":0.1580247,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02690454725407487,"score_gpt":0.2601451152743218,"score_spread":0.233240568020247,"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."}}