{"id":"W2787171718","doi":"10.2218/ijdc.v13i1.620","title":"Data Mining Research with In-copyright and Use-limited Text Datasets: Preliminary Findings from a Systematic Literature Review and Stakeholder Interviews","year":2018,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Library Collection Development and Digital Resources","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Institute of Museum and Library Services","keywords":"Intellectual property; Digitization; Harmonization; Standardization; Scholarly communication; Publishing; Stakeholder; Public relations; Data science; Best practice; Metadata; Political science; Knowledge management; Computer science; Library science; World Wide Web","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.1694068,0.0009641079,0.002896364,0.02423156,0.003543887,0.006029914,0.002567613,0.002869343,0.002750276],"category_scores_gemma":[0.2498991,0.001098781,0.001826102,0.03398568,0.004193028,0.01146683,0.005825837,0.001675594,0.0006030868],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007822537,"about_ca_system_score_gemma":0.03957597,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00499158,"about_ca_topic_score_gemma":0.01827965,"domain_scores_codex":[0.8737494,0.07712789,0.02601455,0.004913446,0.01553042,0.002664228],"domain_scores_gemma":[0.5115516,0.4276288,0.01912918,0.007669001,0.0320187,0.002002737],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"systematic_review","study_design_scores_codex":[0.0004350156,0.0004374702,0.02463302,0.2887212,0.0006986316,0.003892578,0.3930746,0.0007431116,0.008438626,0.008895011,0.006664918,0.2633658],"study_design_scores_gemma":[0.0002133233,0.0006618525,0.02139086,0.2683161,0.002456046,0.001806854,0.5146836,0.000943987,0.007235236,0.00991383,0.1721728,0.0002055747],"study_design_candidate":"systematic_review","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5547795,0.2213556,0.09729534,0.03245011,0.0009212919,0.05354169,0.01599824,0.0002823627,0.02337584],"genre_scores_gemma":[0.6713642,0.1138095,0.1331757,0.007241928,0.000238987,0.0652506,0.006155294,0.0001876994,0.002576279],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9939701,"threshold_uncertainty_score":0.8959196,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1379592423887494,"score_gpt":0.334114570782062,"score_spread":0.1961553283933126,"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."}}