{"id":"W2406594196","doi":"10.5121/csit.2016.60616","title":"A Text Mining Research Based on LDA Topic Modelling","year":2016,"lang":"en","type":"article","venue":"","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":174,"is_retracted":false,"has_abstract":true,"ca_institutions":"Acadia University","funders":"","keywords":"Computer science; Topic model; Natural language processing; Artificial intelligence; Information retrieval","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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002238237,0.0000367315,0.00006477089,0.0001367541,0.0003820228,0.00004248943,0.0001337015,0.0000323588,0.001480689],"category_scores_gemma":[0.0003025434,0.00002253636,0.00004867057,0.0003778314,0.0001225147,0.00005900429,0.0000173696,0.00004689607,0.0001330482],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005991804,"about_ca_system_score_gemma":0.0001244484,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007231758,"about_ca_topic_score_gemma":0.0003207025,"domain_scores_codex":[0.9985658,0.000452413,0.00008966259,0.0001557801,0.000531271,0.0002051028],"domain_scores_gemma":[0.9977511,0.001933754,0.00001525438,0.0000894215,0.0001411879,0.0000692808],"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":[0.00002221646,0.00007564839,0.002919797,0.000003255185,0.00001511378,0.0000045373,0.002801353,0.01268139,0.0002029692,0.4375475,0.002189453,0.5415367],"study_design_scores_gemma":[0.0005454011,0.0001052266,0.001399702,0.00009941353,0.00001148017,1.126349e-7,0.002767607,0.6384366,0.0004070286,0.1501144,0.2058404,0.0002725761],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06962962,0.00001361895,0.4955644,0.0155115,0.00008399764,0.00007203109,4.53643e-7,0.00005234391,0.4190721],"genre_scores_gemma":[0.9451762,0.000005051319,0.02447445,0.0001618968,0.0001848316,0.000005045871,2.611065e-7,0.00000295353,0.02998924],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8755466,"threshold_uncertainty_score":0.9994321,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3539824256636802,"score_gpt":0.5065979207347818,"score_spread":0.1526154950711016,"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."}}