{"id":"W4389207051","doi":"10.22215/etd/2023-15787","title":"Topic Modeling for Intellectual Property Research: Comparing Methods Through Simulation and Application","year":2023,"lang":"en","type":"dissertation","venue":"","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"","keywords":"Trademark; Coherence (philosophical gambling strategy); Intellectual property; Computer science; Metric (unit); Data science; Topic model; Property (philosophy); Cluster analysis; Information retrieval; Data mining; Artificial intelligence; Engineering; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004258283,0.000126775,0.0002825957,0.0002298132,0.001056275,0.0001598417,0.0001774294,0.0002050672,0.00003697521],"category_scores_gemma":[0.00269371,0.00009871741,0.00009752376,0.0006962364,0.00006529205,0.0001482633,0.00003657725,0.0002354519,0.00001452067],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001030765,"about_ca_system_score_gemma":0.0001878798,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005054384,"about_ca_topic_score_gemma":0.00804062,"domain_scores_codex":[0.9978092,0.0007189798,0.000330892,0.0004274381,0.000453248,0.0002601971],"domain_scores_gemma":[0.9948875,0.004102225,0.00008527384,0.0001189778,0.0007500695,0.00005599244],"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":[0.00009529709,0.00004009297,0.00004364334,0.0002014296,0.0001130145,1.524567e-7,0.08374291,0.139599,0.0001259183,0.04925966,0.0003401378,0.7264388],"study_design_scores_gemma":[0.00007831801,0.00001494945,0.00003656571,0.0000300998,0.00003936598,2.643676e-8,0.01539743,0.8655949,0.00003492958,0.1030808,0.01556388,0.0001287604],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0134793,0.0001766563,0.9600317,0.0002396994,0.0002459518,0.0009162598,0.000001048884,0.000137938,0.02477141],"genre_scores_gemma":[0.6689557,0.000233748,0.2388568,0.00003841136,0.0008291089,0.0004858857,0.0006849249,0.00005537468,0.08986001],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.72631,"threshold_uncertainty_score":0.8124123,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.594168270023047,"score_gpt":0.6258136252021957,"score_spread":0.03164535517914879,"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."}}