{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.02790857,0.0010043,0.001391472,0.003283136,0.001149696,0.002919987,0.001941805,0.001974829,0.001812962],"category_scores_gemma":[0.09456863,0.0005062463,0.001522585,0.003380453,0.001234759,0.004325544,0.001745001,0.002657917,0.0004445698],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003557184,"about_ca_system_score_gemma":0.002256005,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01620475,"about_ca_topic_score_gemma":0.01302068,"domain_scores_codex":[0.9897177,0.008153459,0.0003774659,0.0007424156,0.0008132278,0.0001956987],"domain_scores_gemma":[0.7521592,0.2345431,0.002993728,0.005749892,0.003901915,0.0006522017],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007848972,0.0005130008,0.02422627,0.0007594901,0.0006644155,0.00007117313,0.001561789,0.8235736,0.0007116327,0.04576355,0.003533363,0.0978369],"study_design_scores_gemma":[0.00006185078,0.0001285636,0.001314708,0.00006988096,0.00004716168,0.00002081319,0.0002592038,0.9779094,0.0002938845,0.01869021,0.001181826,0.00002239062],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4187989,0.007102898,0.5565653,0.004449736,0.0004010963,0.001007429,0.0009442912,0.0009992478,0.009731166],"genre_scores_gemma":[0.7586234,0.003013774,0.2339766,0.0003705107,0.0002229626,0.001211829,0.0011163,0.0002422211,0.001222388],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9720914,"threshold_uncertainty_score":0.1475964,"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."}}