{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001052477,0.00006868067,0.0000844618,0.00003942445,0.001166424,0.00005922674,0.00009552771,0.00003381804,0.000007107458],"category_scores_gemma":[0.00004220144,0.00004630246,0.00001863512,0.0001309424,0.0001539911,0.0002359083,0.00002819802,0.00008307253,0.000008438007],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008620962,"about_ca_system_score_gemma":0.00002759397,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001316824,"about_ca_topic_score_gemma":0.001102298,"domain_scores_codex":[0.9992008,0.00004919985,0.0001295851,0.0000655385,0.0002364663,0.0003183927],"domain_scores_gemma":[0.9996349,0.00003894604,0.00007276775,0.0001062196,0.00007701394,0.000070174],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0000141847,0.00002516189,0.03041678,0.00004187264,0.00007781749,7.471589e-7,0.779914,0.0000297195,0.000006061139,0.1265885,0.0002090875,0.06267615],"study_design_scores_gemma":[0.0001721737,0.00002532686,0.0004926097,0.00002572173,0.00002770558,0.000002960828,0.04574739,0.0001739677,6.555803e-7,0.0004059834,0.9527766,0.0001489054],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6712188,0.005578338,0.007995305,0.007850893,0.002477061,0.001963564,0.000001417592,0.0008059501,0.3021087],"genre_scores_gemma":[0.9943608,0.000009960403,0.001741819,0.00004966882,0.0004090313,0.00003215992,5.670772e-7,0.000005968271,0.003390023],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9525675,"threshold_uncertainty_score":0.897131,"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."}}