{"id":"W4412609322","doi":"10.16995/dscn.18897","title":"Digital Text Analysis and Early Shakespeare Bibliography: Using Voyant Tools with Bad OCR","year":2025,"lang":"fr","type":"article","venue":"Digital Studies / Le champ numérique","topic":"Library Science and Information Systems","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"St. Francis Xavier University","funders":"","keywords":"Business","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","bibliometrics","scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0003125738,0.0004761141,0.0008063794,0.003757368,0.0006050926,0.009345327,0.0007848212,0.0001337985,0.000006754981],"category_scores_gemma":[0.00008453622,0.0003861537,0.000324002,0.02400217,0.0008086811,0.02844563,0.0009548436,0.0002236623,0.00003347531],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000529308,"about_ca_system_score_gemma":0.0002511677,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002213726,"about_ca_topic_score_gemma":0.00005316963,"domain_scores_codex":[0.997112,0.00006140429,0.0007735712,0.0007823469,0.0005590498,0.0007116618],"domain_scores_gemma":[0.9980748,0.000251846,0.0003196039,0.0006244598,0.00050254,0.0002267355],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"qualitative","study_design_scores_codex":[0.00004828984,0.0002219804,0.7079611,0.0003863212,0.004346339,0.000110547,0.2054933,0.0008921689,0.00001316806,0.002303267,0.00676525,0.07145818],"study_design_scores_gemma":[0.001859037,0.0008563875,0.1570388,0.001950182,0.0000119501,0.0001625845,0.6034563,0.02510967,0.0004156777,0.003597525,0.2032296,0.002312259],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8191295,0.01338447,0.1125539,0.00466406,0.0007690369,0.0005752065,0.0002783309,0.00018525,0.04846033],"genre_scores_gemma":[0.9910196,0.001170707,0.0001285801,0.0004733183,0.0001078294,0.00002331617,0.00002028613,0.00001423968,0.007042123],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5509223,"threshold_uncertainty_score":0.999859,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03526317689244406,"score_gpt":0.2607859084692743,"score_spread":0.2255227315768302,"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."}}