{"id":"W4379157432","doi":"10.1007/s11053-023-10216-1","title":"Applications of Natural Language Processing to Geoscience Text Data and Prospectivity Modeling","year":2023,"lang":"en","type":"article","venue":"Natural Resources Research","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"ca_institutions":"Geological Survey of Canada; Natural Resources Canada","funders":"Natural Resources Canada","keywords":"Prospectivity mapping; Mineral exploration; Computer science; Natural language processing; Naive Bayes classifier; Artificial intelligence; Word (group theory); Geology; Geophysics; Geomorphology; Mathematics; Support vector machine; Structural basin","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.004974508,0.0007836138,0.0006095021,0.005075253,0.0006964875,0.002765342,0.001125546,0.000802706,0.002742196],"category_scores_gemma":[0.02327162,0.0003197726,0.001174971,0.004139702,0.001044766,0.002847849,0.001021549,0.00100083,0.0008215573],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001876142,"about_ca_system_score_gemma":0.001491867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01292848,"about_ca_topic_score_gemma":0.008677105,"domain_scores_codex":[0.9963316,0.00190521,0.0004323502,0.0006117914,0.0006468458,0.00007223753],"domain_scores_gemma":[0.9717899,0.02412177,0.001588594,0.0008359167,0.001483589,0.0001801195],"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":[0.0003140329,0.0005807356,0.02428837,0.00117687,0.0003718431,0.001276763,0.001533445,0.3685425,0.006759564,0.05190378,0.01441442,0.5288377],"study_design_scores_gemma":[0.00001304667,0.00002910114,0.002146974,0.00004837471,0.00002213963,0.0001114648,0.0002910296,0.9436814,0.001573098,0.04754595,0.004513147,0.00002426259],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06522495,0.0008020481,0.9133931,0.004430524,0.0001192685,0.0005278703,0.005550353,0.005237312,0.004714602],"genre_scores_gemma":[0.4371292,0.0005319548,0.5549291,0.0004297275,0.0001927016,0.0004284444,0.005155891,0.0001681773,0.001034742],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01292848,"threshold_uncertainty_score":0.02630806,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06506131511485803,"score_gpt":0.3859891107978926,"score_spread":0.3209277956830345,"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."}}