{"meta":{"query_hash":"7de1237161a1","filters":{"venue":"International Journal of Digital Curation"},"cohort_total":17,"direct_labels_cover":0,"predictions_cover":17,"exported":17,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/7de1237161a1","api":"https://metacan.xera.ac/api/v1/cohort?venue=International+Journal+of+Digital+Curation"},"results":[{"id":"W1485177226","doi":"10.2218/ijdc.v10i1.367","title":"Harmonizing the Metadata Among Diverse Climate Change Datasets","year":2015,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Metadata; Discoverability; Computer science; Harmonization; Raw data; Data science; Data integration; Earth science; Metadata repository; Information retrieval; Database; World Wide Web; Geology","score_opus":0.19513323053149115,"score_gpt":0.3230249752385704,"score_spread":0.12789174470707923,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1485177226","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.29377112,0.0008921574,0.6631266,0.027933663,0.009789086,0.00032234282,0.00033283955,0.00011383988,0.0037183287],"genre_scores_gemma":[0.9980554,0.000036900536,0.0012468997,0.00024488897,0.0003621326,0.000001471667,0.00003655201,0.0000027458254,0.0000130192875],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.99894446,0.000027655298,0.00026294013,0.00009276379,0.0005726622,0.0000994881],"domain_scores_gemma":[0.9990268,0.00007274091,0.00031889896,0.00016873135,0.00035540145,0.00005747317],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00035590882,0.00006843102,0.00008251154,0.00008832404,0.000042448017,0.0012238604,0.0012387125,0.000020128478,0.0000018698016],"category_scores_gemma":[0.0003043357,0.00004321101,0.00005281281,0.00008334876,0.00004073974,0.016329475,0.00033405732,0.00008941603,0.00001832423],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002921365,0.00052594056,0.09365239,0.000015983207,0.0007298342,0.0013152645,0.009456521,0.000478838,0.00036986312,0.3326297,0.044161655,0.51637185],"study_design_scores_gemma":[0.012171429,0.0021929364,0.35672075,0.0009783364,0.0003975027,0.007026685,0.015209766,0.1672948,0.009137831,0.19223858,0.23413283,0.002498572],"about_ca_topic_score_codex":0.0000148602985,"about_ca_topic_score_gemma":0.000012044274,"teacher_disagreement_score":0.70428425,"about_ca_system_score_codex":0.00004434126,"about_ca_system_score_gemma":0.00003797854,"threshold_uncertainty_score":0.99981296},"labels":[],"label_agreement":null},{"id":"W2009293775","doi":"10.2218/ijdc.v8i1.237","title":"A Digital Archives Framework for the Preservation of Cultural Artifacts with Technological Components","year":2013,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Music and Audio Processing","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; McGill University; Centre for Interdisciplinary Research in Music Media and Technology","funders":"","keywords":"Embodied cognition; Computer science; Performative utterance; Readability; Digital preservation; Digitization; Intelligibility (philosophy); World Wide Web; Data science; Knowledge management; Aesthetics; Epistemology","score_opus":0.03866406578836133,"score_gpt":0.2793536446947726,"score_spread":0.24068957890641124,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2009293775","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.13126016,0.000029904055,0.8611611,0.006831286,0.00013220008,0.00011651532,0.0000044247713,0.000013256168,0.00045115603],"genre_scores_gemma":[0.9793572,0.0000043934315,0.020384612,0.000083570005,0.00010164749,0.0000071958,0.0000070974306,0.000003113811,0.000051216688],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99912673,0.000006916192,0.00030665455,0.00008540029,0.0003940876,0.00008019448],"domain_scores_gemma":[0.99859494,0.00031730332,0.00046927299,0.00008526403,0.00050516514,0.000028028058],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00005670688,0.000070966846,0.00009030723,0.000068768895,0.000046006164,0.0008884093,0.0006526609,0.00003011416,0.0000039792612],"category_scores_gemma":[0.0003311136,0.00003772901,0.000062961524,0.0000909624,0.00009931672,0.0074645355,0.0000741115,0.0001044164,0.0000026856503],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018480454,0.0002983158,0.0052257874,0.000020039937,0.00024248022,0.0000055441014,0.0012720543,0.0017214331,0.0049137827,0.17292331,0.0004635972,0.8127289],"study_design_scores_gemma":[0.0018021297,0.0010221201,0.035955273,0.00082792033,0.000030543237,0.0004017873,0.0009278337,0.12949121,0.016267445,0.80586416,0.0069962856,0.00041328487],"about_ca_topic_score_codex":7.0464847e-7,"about_ca_topic_score_gemma":1.5627077e-7,"teacher_disagreement_score":0.848097,"about_ca_system_score_codex":0.000015387972,"about_ca_system_score_gemma":0.000034217235,"threshold_uncertainty_score":0.8566952},"labels":[],"label_agreement":null},{"id":"W2062133553","doi":"10.2218/ijdc.v7i2.227","title":"Digital Forensics Formats: Seeking a Digital Preservation Storage Container Format for Web Archiving","year":2012,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"European Commission","keywords":"Computer science; Container (type theory); File format; Digital forensics; Digital preservation; Image file formats; Database; World Wide Web; Point (geometry); File system; Information retrieval; Operating system; Engineering","score_opus":0.12336317467511576,"score_gpt":0.3905250777087929,"score_spread":0.26716190303367715,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2062133553","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1350752,0.000047290905,0.84658235,0.0018159974,0.0020456875,0.00031546544,0.00089904177,0.000029012486,0.013189931],"genre_scores_gemma":[0.99690014,0.0000048821976,0.0007112989,0.00023130629,0.0009000104,0.000008231188,0.00039771924,0.000014970725,0.00083144923],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.996238,0.00002780698,0.0013357528,0.00015904842,0.0019580878,0.00028133485],"domain_scores_gemma":[0.99619645,0.00091118074,0.0012060009,0.00020428539,0.001331589,0.00015049685],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0014325152,0.00016638881,0.00023944616,0.0004307195,0.000098982055,0.0046199416,0.00088070804,0.00004984181,0.00001971596],"category_scores_gemma":[0.00491031,0.00012909323,0.00026675034,0.00018698181,0.000056257402,0.06410739,0.00027363535,0.00013273193,0.000051267496],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009819688,0.0005622851,0.026592674,0.00003762144,0.0004311036,0.000014035015,0.0031256948,0.0010254909,0.000343414,0.08937727,0.03977892,0.8377295],"study_design_scores_gemma":[0.0029799123,0.00049684366,0.007265578,0.00019265903,0.00004934685,0.0002907053,0.0052358247,0.03358188,0.00018564377,0.1719073,0.77725714,0.0005571709],"about_ca_topic_score_codex":7.7961937e-7,"about_ca_topic_score_gemma":0.0000038497296,"teacher_disagreement_score":0.8618249,"about_ca_system_score_codex":0.00015945514,"about_ca_system_score_gemma":0.00007789496,"threshold_uncertainty_score":0.99641335},"labels":[],"label_agreement":null},{"id":"W2128634280","doi":"10.2218/ijdc.v8i1.238","title":"The Critical Role of Institutional Services in Open Access Advocacy","year":2013,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Service (business); Scholarly communication; Rhetorical question; Rhetoric; Open data; World Wide Web; Public relations; Open science; Computer science; Discipline; Open access journal; Internet privacy; Political science; Business; Publishing; Marketing; MEDLINE","score_opus":0.07802627827890123,"score_gpt":0.44013263238195005,"score_spread":0.3621063541030488,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2128634280","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.56367284,0.0006064359,0.29395005,0.0813608,0.0023568308,0.0009047651,0.000031904565,0.000025017544,0.057091333],"genre_scores_gemma":[0.99860704,0.00010939545,0.0010517841,0.00008367392,0.00008227432,0.000007817279,0.0000071130353,0.0000021762432,0.000048754522],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99845254,0.00005329136,0.00045868263,0.00010970313,0.00082272023,0.000103061255],"domain_scores_gemma":[0.99809426,0.00042487832,0.00032915807,0.00017585009,0.00092879124,0.00004704631],"candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00052826916,0.00005562924,0.0000775446,0.00019025737,0.000054017128,0.015113561,0.0065945955,0.000016984028,0.000015432806],"category_scores_gemma":[0.0009278279,0.000038597074,0.00003418021,0.00019605897,0.00006745785,0.21740091,0.0017943075,0.0001240497,0.000016062986],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00006694965,0.00025926222,0.010121954,0.000010727058,0.00006853124,0.000023430115,0.000253091,0.000439971,0.0005102443,0.88144904,0.0002313091,0.10656549],"study_design_scores_gemma":[0.0020129168,0.0004093834,0.2604335,0.0004235362,0.000011759318,0.0002457974,0.0012553739,0.06247905,0.00235913,0.5782036,0.0918323,0.00033364975],"about_ca_topic_score_codex":0.0001381835,"about_ca_topic_score_gemma":0.000035083813,"teacher_disagreement_score":0.43493414,"about_ca_system_score_codex":0.000063556705,"about_ca_system_score_gemma":0.00019918695,"threshold_uncertainty_score":0.9987802},"labels":[],"label_agreement":null},{"id":"W2160348740","doi":"10.2218/ijdc.v6i1.176","title":"Towards a Holistic Approach to Policy Interoperability in Digital Libraries and Digital Repositories","year":2011,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"U.S. Geological Survey","keywords":"Interoperability; Digital library; Computer science; World Wide Web; Metadata; Digital transformation; Digital preservation; Reuse; Key (lock); Digital curation; Domain (mathematical analysis); Cross-domain interoperability; Semantic interoperability; Computer security; Engineering","score_opus":0.20938298116014792,"score_gpt":0.3937515392860538,"score_spread":0.18436855812590586,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2160348740","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.7927015,0.00004452895,0.10077623,0.002276351,0.0011419726,0.00021955787,0.00034618186,0.000023680881,0.10247002],"genre_scores_gemma":[0.9983625,0.000003527225,0.00073060655,0.00013381563,0.00028285643,0.0000032198482,0.00003036416,0.0000068364866,0.00044629595],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9973471,0.000031599557,0.0010428344,0.00026410312,0.0011793212,0.00013505116],"domain_scores_gemma":[0.99842834,0.0001747988,0.0003768324,0.00020597198,0.0006760016,0.0001380237],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0005932123,0.00013106281,0.00022619203,0.00077837496,0.000034675297,0.0050131734,0.0008533098,0.000040377745,0.000011779972],"category_scores_gemma":[0.007239138,0.000097511234,0.000085999076,0.00045283066,0.00013641735,0.01836336,0.00046455025,0.00012247587,0.000023999879],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.001508245,0.0015860363,0.078019544,0.000020972368,0.00019958099,0.00011254336,0.020616239,0.00009767021,0.00004151585,0.45861188,0.00290976,0.436276],"study_design_scores_gemma":[0.0013235633,0.0008435978,0.120181836,0.00012942248,0.000015636751,0.00046568265,0.011408739,0.0004534826,0.00032526843,0.8345644,0.029794952,0.00049344497],"about_ca_topic_score_codex":0.00002830951,"about_ca_topic_score_gemma":0.000005439092,"teacher_disagreement_score":0.43578255,"about_ca_system_score_codex":0.00010771673,"about_ca_system_score_gemma":0.00013173108,"threshold_uncertainty_score":0.9960197},"labels":[],"label_agreement":null},{"id":"W2543266858","doi":"10.2218/ijdc.v11i1.411","title":"Towards a Collaborative National Research Data Management Network","year":2016,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Queen's University; Canadian Association of Research Libraries; University of Alberta","funders":"","keywords":"RDM; Context (archaeology); Deliverable; Knowledge management; Computer science; Corporate governance; Process management; Data governance; Government (linguistics); Open data; Data sharing; Business; World Wide Web; Data quality; Engineering; Service (business); Systems engineering","score_opus":0.26605477904163305,"score_gpt":0.47887570378532124,"score_spread":0.2128209247436882,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2543266858","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0006446745,0.00010637078,0.9105577,0.034176208,0.0010806413,0.00020711243,0.0001240185,0.000024759029,0.053078517],"genre_scores_gemma":[0.9698119,0.0009697503,0.026531521,0.00015613745,0.0011331581,0.000009606581,0.000060321272,0.000010425394,0.0013171562],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99423826,0.00018291046,0.00048751628,0.0003200232,0.0045354497,0.00023584311],"domain_scores_gemma":[0.99334115,0.00050096965,0.000403556,0.0005132338,0.005146346,0.00009475846],"candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00454219,0.00009194578,0.000101227524,0.00048304486,0.00008772271,0.0044145547,0.0058188424,0.000028769911,0.000020318997],"category_scores_gemma":[0.0028386747,0.00006276766,0.000036803565,0.0006483699,0.00008133906,0.11681487,0.0023803394,0.00015871419,0.00007698323],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011066936,0.00013579735,0.00021527184,0.0000044829094,0.0003167695,0.0001474542,0.00006601493,0.00010136851,0.00008585624,0.8131982,0.044193648,0.14142445],"study_design_scores_gemma":[0.0014213606,0.00030000496,0.0022501813,0.00022000867,0.000009250266,0.000115676325,0.00015129549,0.0026829203,0.00015065596,0.18622598,0.80627024,0.00020240794],"about_ca_topic_score_codex":0.000002325022,"about_ca_topic_score_gemma":0.0000055403775,"teacher_disagreement_score":0.96916723,"about_ca_system_score_codex":0.0003833018,"about_ca_system_score_gemma":0.00043809408,"threshold_uncertainty_score":0.9995602},"labels":[],"label_agreement":null},{"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"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","score_opus":0.13795924238874943,"score_gpt":0.334114570782062,"score_spread":0.1961553283933126,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2787171718","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.85766566,0.094394445,0.02801798,0.013126815,0.0010128311,0.0018523003,0.0014159577,0.000087429195,0.0024265947],"genre_scores_gemma":[0.99201864,0.0024486796,0.004285991,0.0004552272,0.00012334542,0.000007239734,0.00040709195,0.000010413653,0.00024337773],"study_design_codex":"not_applicable","study_design_gemma":"systematic_review","domain_scores_codex":[0.998363,0.000094129486,0.0005499436,0.00026776516,0.00060910673,0.00011606314],"domain_scores_gemma":[0.99853307,0.00041749343,0.00026355757,0.00025832845,0.00045238796,0.000075141295],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00051325787,0.00011567751,0.00022067985,0.00044794357,0.00006743157,0.0031430942,0.00088924286,0.00003614997,0.00000780283],"category_scores_gemma":[0.0006213369,0.000078348836,0.00001565551,0.00051961385,0.00007643674,0.02182724,0.0005649068,0.0001622977,0.0000036711122],"study_design_candidate":"systematic_review","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0049850945,0.0024458237,0.1738453,0.03248719,0.0038593363,0.0037746853,0.073157735,0.00001240348,0.00060547236,0.014198856,0.5286511,0.16197701],"study_design_scores_gemma":[0.01174595,0.008673862,0.12785308,0.6507705,0.00040134584,0.012132478,0.0036611762,0.0442048,0.0012941937,0.013854429,0.12232898,0.0030792255],"about_ca_topic_score_codex":0.0000019989243,"about_ca_topic_score_gemma":0.000012976523,"teacher_disagreement_score":0.6182833,"about_ca_system_score_codex":0.000041347263,"about_ca_system_score_gemma":0.00008038036,"threshold_uncertainty_score":0.9978917},"labels":[],"label_agreement":null},{"id":"W2804583692","doi":"10.2218/ijdc.v12i2.579","title":"Mobilising a Nation: RDM Training and Education in South Africa","year":2018,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian Society of Intestinal Research","funders":"","keywords":"RDM; Digital curation; Research data; Service (business); Data curation; Library science; Subject (documents); Political science; Public relations; Sociology; World Wide Web; Business; Computer science; Pedagogy; Marketing","score_opus":0.12151935611337122,"score_gpt":0.38094310559413974,"score_spread":0.2594237494807685,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2804583692","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.32023147,0.00020825677,0.6522962,0.005397811,0.0014749283,0.00016272093,0.000005827171,0.000023033264,0.02019974],"genre_scores_gemma":[0.9938837,0.00002202023,0.0054776524,0.00005701129,0.00035667708,0.0000017894599,0.0000048981506,0.00000316677,0.00019304595],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9988668,0.000036695554,0.0003114852,0.000133943,0.00056113105,0.000089923284],"domain_scores_gemma":[0.99883264,0.00010001127,0.0003113244,0.00010352306,0.00060585066,0.000046630415],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0005261014,0.00005624805,0.000066260014,0.00045244448,0.000033912944,0.002919805,0.00064942695,0.000019169403,0.000004082776],"category_scores_gemma":[0.0008836632,0.000052187817,0.000021278453,0.00020956997,0.000040670642,0.059431326,0.0001511684,0.00009796402,0.0000057408956],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000093190254,0.00032884712,0.004989408,0.0000112590205,0.00008990543,0.00004218162,0.023530684,0.00009584124,0.00078382704,0.24332576,0.0003736352,0.72633547],"study_design_scores_gemma":[0.0077685183,0.0037786772,0.08252467,0.0017048669,0.00006026059,0.001767913,0.02597853,0.14241232,0.0027070479,0.39749035,0.3320605,0.0017463485],"about_ca_topic_score_codex":0.0000020127777,"about_ca_topic_score_gemma":0.0000049262544,"teacher_disagreement_score":0.7245891,"about_ca_system_score_codex":0.00008510417,"about_ca_system_score_gemma":0.0002540314,"threshold_uncertainty_score":0.99811524},"labels":[],"label_agreement":null},{"id":"W2805287170","doi":"10.2218/ijdc.v12i2.583","title":"The Changing Influence of Journal Data Sharing Policies on Local RDM Practices","year":2018,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Scopus; Citation; RDM; Scripting language; Science policy; Library science; Impact factor; Computer science; Python (programming language); Publishing; Data sharing; Order (exchange); Political science; Business; Public administration; MEDLINE; Medicine; Finance; Law","score_opus":0.17890550097226754,"score_gpt":0.44477389649190174,"score_spread":0.2658683955196342,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2805287170","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.30117807,0.00013803047,0.67377496,0.019912599,0.0013236912,0.00010021786,0.000021002852,0.000015056916,0.0035363801],"genre_scores_gemma":[0.9975991,0.00035137514,0.001136785,0.00012689718,0.0006736573,5.08771e-7,0.0000058183746,0.0000047202616,0.00010111726],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9975369,0.00006255422,0.00054589816,0.00017156506,0.0015064087,0.00017668356],"domain_scores_gemma":[0.99507636,0.00061438733,0.00218692,0.0005567356,0.0015002684,0.0000653479],"candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0023323435,0.00008077871,0.000089690846,0.00038517985,0.00019263715,0.0071524573,0.0067351605,0.000020030648,0.0000027886815],"category_scores_gemma":[0.006176362,0.000054267766,0.00004073435,0.00027729536,0.00014910234,0.105940826,0.0015754343,0.00025617713,0.000012874054],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007260848,0.00046117458,0.007108179,0.000026749794,0.0009812396,0.00021450702,0.0032176694,0.005444121,0.002058018,0.6978232,0.0027087862,0.27923027],"study_design_scores_gemma":[0.0044477684,0.0066767368,0.05446332,0.0026257304,0.00015148646,0.006269163,0.009401717,0.3159764,0.015482117,0.11062148,0.47258988,0.0012941894],"about_ca_topic_score_codex":0.000011331511,"about_ca_topic_score_gemma":0.000011171755,"teacher_disagreement_score":0.696421,"about_ca_system_score_codex":0.000091640526,"about_ca_system_score_gemma":0.00014293374,"threshold_uncertainty_score":0.99863887},"labels":[],"label_agreement":null},{"id":"W2907551312","doi":"10.2218/ijdc.v13i1.534","title":"Participatory Prototype Design: Developing a Sustainable Metadata Curation Workflow for Maternal Child Health Research","year":2018,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Faculty of Medicine and Dentistry, University of Alberta; Women and Children's Health Research Institute; University of Alberta; Harvard University; Purdue University; Andrew W. Mellon Foundation","keywords":"Workflow; Metadata; Data curation; Citizen journalism; Computer science; Digital curation; Knowledge management; Participatory action research; Participatory design; World Wide Web; Process management; Business; Sociology; Database; Engineering; Operations management","score_opus":0.41671221442863426,"score_gpt":0.4953847330675692,"score_spread":0.07867251863893493,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2907551312","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0019411611,0.00008383095,0.9756614,0.020476174,0.00049260637,0.00085382006,0.00000979977,0.000023718085,0.00045750773],"genre_scores_gemma":[0.9209918,0.00010637245,0.076971985,0.00024352084,0.0008918873,0.00009050366,0.0000349095,0.000014914725,0.00065415114],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99670845,0.00029567405,0.00067519135,0.00028076916,0.0016062876,0.00043363345],"domain_scores_gemma":[0.99471885,0.0003701156,0.00062678714,0.00030563588,0.0038484372,0.0001301862],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0053181085,0.00011333845,0.00015044856,0.0006069462,0.00031998346,0.009114065,0.002117416,0.000035175566,0.0000064081123],"category_scores_gemma":[0.0028752084,0.00009909035,0.00005861997,0.0004428013,0.00008697601,0.12841097,0.00047871398,0.00022940275,0.00001788916],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009085607,0.00022357538,0.00047251023,0.000077414625,0.00030378698,0.000070965914,0.0008160279,0.00022919326,0.00013323798,0.9210519,0.0041986275,0.07151419],"study_design_scores_gemma":[0.0039060966,0.007570234,0.0047175162,0.0010446468,0.00003113143,0.0005195196,0.0010911927,0.061705053,0.011387458,0.24688368,0.660353,0.0007904639],"about_ca_topic_score_codex":0.000016091804,"about_ca_topic_score_gemma":0.0000101459045,"teacher_disagreement_score":0.9190506,"about_ca_system_score_codex":0.00044854518,"about_ca_system_score_gemma":0.0008007086,"threshold_uncertainty_score":0.9919146},"labels":[],"label_agreement":null},{"id":"W2973161955","doi":"10.2218/ijdc.v14i1.586","title":"Identifying Topical Coverages of Curricula using Topic Modeling and Visualization Techniques: A Case of Digital and Data Curation","year":2019,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Digital curation; Data curation; Syllabus; Computer science; Curriculum; Data science; Context (archaeology); Latent Dirichlet allocation; Metadata; Topic model; Cyberinfrastructure; World Wide Web; Visualization; Data management; Information retrieval; Database; Data mining","score_opus":0.14827183941897354,"score_gpt":0.4454172668164202,"score_spread":0.29714542739744665,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2973161955","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.43447393,0.000046849244,0.5651111,0.00011572372,0.000090070214,0.00007097544,0.000019601344,0.0000043499194,0.00006739702],"genre_scores_gemma":[0.9942589,0.00013823093,0.00548444,0.000007252337,0.000056257093,3.966422e-7,0.000037777263,0.0000036573376,0.000013097211],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99877346,0.0000316191,0.00047484378,0.00017182429,0.00048514723,0.00006312236],"domain_scores_gemma":[0.9986635,0.00009495238,0.0005134094,0.00021217664,0.00048315807,0.00003281716],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0003367566,0.00006719592,0.00011861571,0.0002486134,0.000022832774,0.0016551102,0.00054126646,0.00002822644,0.0000016244243],"category_scores_gemma":[0.00045841828,0.000060901173,0.00002292569,0.00011889857,0.000026890344,0.078771554,0.00056671764,0.00007351315,2.4958464e-7],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00032791635,0.0007914151,0.0653547,0.0006160922,0.0006415649,0.0005911177,0.0023386485,0.0061097257,0.031642783,0.55588394,0.000061419305,0.33564067],"study_design_scores_gemma":[0.00042136197,0.0002146859,0.00024605053,0.0002780725,0.00002072057,0.001114478,0.00028474248,0.9883304,0.001602943,0.0071807774,0.00018686084,0.00011891033],"about_ca_topic_score_codex":0.000014655846,"about_ca_topic_score_gemma":0.0000018491062,"teacher_disagreement_score":0.98222065,"about_ca_system_score_codex":0.000032521177,"about_ca_system_score_gemma":0.000043629603,"threshold_uncertainty_score":0.99938124},"labels":[],"label_agreement":null},{"id":"W3048135048","doi":"10.2218/ijdc.v15i1.728","title":"“You say potato, I say potato†Mapping Digital Preservation and Research Data Management Concepts towards Collective Curation and Preservation Strategies","year":2020,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Data curation; Digital curation; Computer science; Research data; Digital preservation; Data management; Data science; World Wide Web; Database","score_opus":0.24964092585801395,"score_gpt":0.4301739196899755,"score_spread":0.18053299383196153,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3048135048","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.083147794,0.00027408244,0.85787755,0.047661602,0.00032645263,0.00076207024,0.00014450292,0.00005799193,0.009747972],"genre_scores_gemma":[0.9936192,0.000531564,0.004655896,0.00020012796,0.00027476504,0.0000122029915,0.00031960956,0.000012959615,0.000373654],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9963022,0.00019092398,0.00067485095,0.00054454553,0.0020460035,0.00024147575],"domain_scores_gemma":[0.9969652,0.00039330748,0.00056381454,0.00038588786,0.0015217654,0.00017000659],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0013242215,0.0001693222,0.0001793967,0.0004644941,0.00017805958,0.01628489,0.0021600404,0.000052705574,0.0000059406034],"category_scores_gemma":[0.0021809493,0.00015945862,0.00003356275,0.00065485406,0.00012769426,0.25293705,0.0021523724,0.00033537563,0.0000064206097],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000985305,0.0004880335,0.010063379,0.0003247491,0.00136021,0.00032199518,0.012772598,0.0011008773,0.001740124,0.69856286,0.011794454,0.2604854],"study_design_scores_gemma":[0.00686959,0.0024970276,0.10831164,0.0007464839,0.00009236343,0.00030285394,0.015750086,0.43902645,0.0006621486,0.28564373,0.13880183,0.0012958086],"about_ca_topic_score_codex":0.000016280637,"about_ca_topic_score_gemma":0.0000065172217,"teacher_disagreement_score":0.91047144,"about_ca_system_score_codex":0.0001655549,"about_ca_system_score_gemma":0.00028127097,"threshold_uncertainty_score":0.9847363},"labels":[],"label_agreement":null},{"id":"W3116320491","doi":"10.2218/ijdc.v15i1.708","title":"Access Some Areas: Reforming Access Categories for Data in a Social Science Data Archive","year":2020,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Data access; Data science; Computer science; World Wide Web; Database","score_opus":0.34453384877516346,"score_gpt":0.4845747686055644,"score_spread":0.14004091983040096,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3116320491","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0069853878,0.000037853242,0.95657575,0.03363831,0.0005677537,0.00024120108,0.00052517775,0.000019908373,0.0014086604],"genre_scores_gemma":[0.99559706,0.00007230772,0.0026211278,0.00037750375,0.00073671754,0.0000037897628,0.0005750545,0.0000059667764,0.000010451889],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9975631,0.00003143541,0.00048535888,0.00042071077,0.001309126,0.00019023319],"domain_scores_gemma":[0.9980146,0.00022954901,0.0005876312,0.00053674116,0.0005396042,0.000091836424],"candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.0011900383,0.00008861253,0.00012306431,0.0003956904,0.00011193532,0.014726032,0.02184268,0.000016686234,0.0000017376219],"category_scores_gemma":[0.004447126,0.0000771525,0.00002512528,0.0004894197,0.00010799246,0.44720206,0.009225845,0.00017536848,0.0000024056794],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00069279585,0.0003918023,0.009961085,0.00009459296,0.00025726418,0.00023603381,0.0018770462,0.00026026598,0.0011241552,0.7704402,0.005673778,0.20899093],"study_design_scores_gemma":[0.0037829112,0.0005267083,0.01864703,0.00022197624,0.000042556178,0.00016163009,0.0010064706,0.62509465,0.0015145306,0.25740144,0.09081923,0.0007808983],"about_ca_topic_score_codex":0.000018373325,"about_ca_topic_score_gemma":0.00001472395,"teacher_disagreement_score":0.9886117,"about_ca_system_score_codex":0.00010569829,"about_ca_system_score_gemma":0.0004960768,"threshold_uncertainty_score":0.99878734},"labels":[],"label_agreement":null},{"id":"W4293077132","doi":"10.2218/ijdc.v16i1.769","title":"Research Data Management Practices at the University of Namibia: Moving Towards Adoption","year":2022,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Mastercard Foundation","keywords":"RDM; Process (computing); Knowledge management; Research data; Sociology; Business; Computer science; Data science; Data curation","score_opus":0.25793408400286144,"score_gpt":0.43282386043997434,"score_spread":0.1748897764371129,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4293077132","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.19144736,0.0006872804,0.6827112,0.07715207,0.002200565,0.0007045335,0.0003201311,0.00004176308,0.044735115],"genre_scores_gemma":[0.995472,0.0005824725,0.0025906544,0.000029095203,0.00007538055,7.9098476e-7,0.000087578854,0.000003560019,0.0011584578],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9965758,0.0003273576,0.00026366467,0.00021037918,0.0025054484,0.000117398944],"domain_scores_gemma":[0.9974008,0.0003473673,0.0009854352,0.00058804627,0.0006431565,0.00003515547],"candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.0038601093,0.000050860024,0.00006487511,0.00032510134,0.00029317068,0.0012598245,0.0058627226,0.000010959441,0.00006235514],"category_scores_gemma":[0.00068614824,0.00004357554,0.000037865746,0.00039674368,0.00006934903,0.055190384,0.007568535,0.00030203612,0.000011203833],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012928976,0.0012474152,0.0028430203,0.00008358276,0.0014848736,0.0012610834,0.0030037342,0.011722673,0.00095304695,0.4266997,0.0355247,0.5138833],"study_design_scores_gemma":[0.0009117887,0.0004337233,0.009693597,0.000051244715,0.000034821496,0.0002761497,0.0080460245,0.03829304,0.000088202425,0.004516514,0.93750936,0.00014553628],"about_ca_topic_score_codex":0.000048786045,"about_ca_topic_score_gemma":0.000026053081,"teacher_disagreement_score":0.90198463,"about_ca_system_score_codex":0.00034853586,"about_ca_system_score_gemma":0.00012864954,"threshold_uncertainty_score":0.99977696},"labels":[],"label_agreement":null},{"id":"W4303415524","doi":"10.2218/ijdc.v17i1.782","title":"OpenStack Swift: An Ideal Bit-Level Object Storage System for Digital Preservation","year":2022,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Advanced Data Storage Technologies","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Alberta Library; University of Alberta","funders":"","keywords":"Computer science; Digital preservation; Swift; Object storage; Obsolescence; Scalability; Cloud computing; Object (grammar); Cloud storage; Computer data storage; Software; Database; Software engineering; Operating system","score_opus":0.05292288886082759,"score_gpt":0.3047834880594124,"score_spread":0.25186059919858483,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4303415524","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.047470216,0.000044502092,0.9482679,0.00095386006,0.001570425,0.00024854546,0.0008368053,0.00015080364,0.00045698602],"genre_scores_gemma":[0.9839847,0.000002112149,0.015280981,0.000050297225,0.00021992734,0.000026926191,0.00031389398,0.000015283727,0.0001058444],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99791276,0.000030348128,0.0006202726,0.0002587674,0.001008553,0.00016930523],"domain_scores_gemma":[0.9979344,0.00013748875,0.0007780249,0.00030642725,0.0007798836,0.000063762214],"candidate_categories":["scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.00031414593,0.00014441794,0.00017530311,0.0003590714,0.00016490054,0.0013378481,0.0021383618,0.000039416445,0.000004801557],"category_scores_gemma":[0.00043141053,0.00014127979,0.00011143617,0.0002501351,0.000033467277,0.027119188,0.00060468016,0.00020626595,0.000004241674],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00076603703,0.0007717328,0.0010136104,0.000047496207,0.00034698076,0.00038384678,0.0014546043,0.097660005,0.0040730936,0.6461819,0.0043529198,0.24294777],"study_design_scores_gemma":[0.009355438,0.007910163,0.0033045083,0.00030584564,0.00006930917,0.0048068627,0.006769148,0.6732277,0.010438215,0.20068918,0.0809722,0.0021513796],"about_ca_topic_score_codex":0.000004112859,"about_ca_topic_score_gemma":0.0000034537131,"teacher_disagreement_score":0.9365145,"about_ca_system_score_codex":0.0005718675,"about_ca_system_score_gemma":0.0001773862,"threshold_uncertainty_score":0.9996989},"labels":[],"label_agreement":null},{"id":"W4414244142","doi":"10.2218/ijdc.v19i1.906","title":"A Maturity Model for Urban Dataset Metadata","year":2025,"lang":"en","type":"article","venue":"International Journal of Digital Curation","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Metadata; Documentation; Maturity (psychological); Capability Maturity Model; Relevance (law); Graph; Plug-in; Software","score_opus":0.2141903317703833,"score_gpt":0.4785366575588719,"score_spread":0.2643463257884886,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4414244142","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.002133158,0.0000491664,0.9800603,0.007966633,0.0009948907,0.000111906156,0.006940398,0.000005273505,0.0017383035],"genre_scores_gemma":[0.98939437,0.000010541612,0.0032160291,0.0010466509,0.00017710976,0.0000041138333,0.0018770165,0.000003425577,0.0042707296],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9979696,0.00002877164,0.0007531368,0.00015143745,0.0010227802,0.00007426737],"domain_scores_gemma":[0.99800473,0.00039788647,0.00045467506,0.00023320355,0.0008715678,0.0000379319],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0014521793,0.00006764754,0.00013923638,0.0003243488,0.000040753286,0.0018568835,0.0012243174,0.00002635149,0.000019363997],"category_scores_gemma":[0.003281104,0.000050745657,0.00010534367,0.00014095228,0.000030039198,0.0098719755,0.0002063727,0.000068941554,0.000016524567],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002582959,0.00015559363,0.0001708632,0.0000046771593,0.00015711982,0.000004612083,0.00008910746,0.0013124911,0.000059772632,0.22445163,0.738119,0.0352168],"study_design_scores_gemma":[0.00065154984,0.000042085692,0.0002039174,0.000025270538,0.000030444338,0.000008720334,0.00018485995,0.06703707,0.00014940422,0.41544506,0.51614624,0.0000753836],"about_ca_topic_score_codex":0.0000020754635,"about_ca_topic_score_gemma":0.000012510139,"teacher_disagreement_score":0.98726124,"about_ca_system_score_codex":0.000051292674,"about_ca_system_score_gemma":0.00010153944,"threshold_uncertainty_score":0.9991793},"labels":[],"label_agreement":null},{"id":"W4417200670","doi":"10.2218/ijdc.v19i1.1050","title":"Informed Consent Contexts in a Multidisciplinary Research Data Repository","year":2025,"lang":"","type":"article","venue":"International Journal of Digital Curation","topic":"Research Data Management Practices","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Mount Royal University","funders":"","keywords":"Informed consent; Metadata; Multidisciplinary approach; Research ethics; Data sharing; Ethos; Information repository; Data collection","score_opus":0.2930769846915159,"score_gpt":0.517679969062208,"score_spread":0.22460298437069215,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417200670","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.21813972,0.0053752023,0.5088231,0.11188337,0.020294582,0.0025437614,0.00054588873,0.000060108378,0.13233423],"genre_scores_gemma":[0.9925222,0.0026587413,0.00095243525,0.00007097457,0.0003341311,0.0000063173547,0.00011810642,0.000007980489,0.0033291078],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.993512,0.0005511292,0.0018783918,0.0005784695,0.0030586042,0.0004214354],"domain_scores_gemma":[0.98929447,0.0049792775,0.0010929302,0.0012725191,0.0032089422,0.00015185888],"candidate_categories":["metaresearch","scholarly_communication","open_science"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.004671011,0.00019729635,0.00028405827,0.002389269,0.00017545826,0.014022082,0.009174661,0.00010322154,0.000009314631],"category_scores_gemma":[0.020959033,0.00018964251,0.00009349984,0.0011097207,0.00033246697,0.16926005,0.0068171592,0.00097302074,0.000027988066],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":true,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0065503684,0.006750537,0.07969769,0.00089788076,0.0029803657,0.009662398,0.0067906603,0.002094403,0.0023986145,0.49120072,0.04473206,0.3462443],"study_design_scores_gemma":[0.023567382,0.0039306707,0.20009683,0.023615625,0.00018156333,0.0022945134,0.011768678,0.35655612,0.00282392,0.07184752,0.30129966,0.002017529],"about_ca_topic_score_codex":0.00009969911,"about_ca_topic_score_gemma":0.00046257046,"teacher_disagreement_score":0.7743825,"about_ca_system_score_codex":0.0008676562,"about_ca_system_score_gemma":0.0057553193,"threshold_uncertainty_score":0.99988115},"labels":[],"label_agreement":null}]}