{"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.10432839,0.002985325,0.82129264,0.007872605,0.000519205,0.0017368473,0.02824094,0.005607885,0.027416274],"genre_scores_gemma":[0.26044407,0.0023260766,0.67632467,0.0014770385,0.00016644842,0.0008366086,0.054004386,0.0018188504,0.002601913],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9770768,0.008729301,0.0037944475,0.0033762392,0.0064936555,0.0005294872],"domain_scores_gemma":[0.93721235,0.013842178,0.0042522578,0.0321063,0.0118941255,0.0006928112],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.039649766,0.0005237095,0.0012155048,0.019103095,0.0019282372,0.0060707196,0.002824248,0.00078431313,0.0012485726],"category_scores_gemma":[0.07154022,0.0007275417,0.0016867869,0.02686192,0.0019427892,0.010642008,0.008324682,0.0020589102,0.000683461],"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":[0.0003325905,0.00046405135,0.1008342,0.0022779421,0.0018720506,0.00070569,0.012231968,0.023382433,0.021263815,0.14700317,0.03223246,0.6573996],"study_design_scores_gemma":[0.000113591406,0.00019718014,0.08952477,0.0021348,0.0011922622,0.00089853274,0.015296384,0.040541425,0.040131275,0.23078577,0.5787541,0.00042996614],"about_ca_topic_score_codex":0.0070150634,"about_ca_topic_score_gemma":0.008859264,"teacher_disagreement_score":0.039649766,"about_ca_system_score_codex":0.0022485382,"about_ca_system_score_gemma":0.0057963817,"threshold_uncertainty_score":0.20969051},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0075808642,0.004654643,0.8628668,0.0048666424,0.00040399053,0.00051744754,0.0007164623,0.0013037679,0.11708947],"genre_scores_gemma":[0.21493964,0.0057698595,0.72999877,0.0009039521,0.00050200376,0.0012031792,0.0016016411,0.00065044774,0.044430457],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9956684,0.0018345907,0.00056489545,0.0006282863,0.0009798696,0.00032392054],"domain_scores_gemma":[0.994618,0.0016184006,0.00054056296,0.0018881145,0.0007788016,0.00055626256],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006108504,0.0009866543,0.0009066897,0.0072818836,0.0052235485,0.020651553,0.004971988,0.0040428727,0.008542029],"category_scores_gemma":[0.006943894,0.0008043952,0.0023139382,0.0074048825,0.012927474,0.0227712,0.010833262,0.004469407,0.0025589946],"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":[0.000010818104,0.000015343854,0.00010598716,0.00010166522,0.000006504616,0.00019887267,0.0017427994,0.00086869724,0.00018235593,0.98720086,0.00077796704,0.0087881815],"study_design_scores_gemma":[0.000026848518,0.00003158732,0.00021275043,0.0004290399,0.000045512643,0.0008199564,0.0030830426,0.006703174,0.0007977529,0.7305316,0.2572713,0.000047532187],"about_ca_topic_score_codex":0.009884418,"about_ca_topic_score_gemma":0.009313523,"teacher_disagreement_score":0.020651553,"about_ca_system_score_codex":0.0051322505,"about_ca_system_score_gemma":0.006882258,"threshold_uncertainty_score":0.037237227},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.2952585,0.0024616714,0.6102983,0.006378303,0.00062811264,0.0034002487,0.004636802,0.0056996443,0.07123853],"genre_scores_gemma":[0.31308085,0.0013613391,0.67361534,0.00062933186,0.00018229982,0.00097421807,0.003152175,0.0009890691,0.00601547],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99224806,0.0026392513,0.0010376077,0.00049582263,0.0031344728,0.00044487067],"domain_scores_gemma":[0.97441804,0.0063884323,0.0027468787,0.0069192383,0.008649086,0.00087834924],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.017635373,0.00069751375,0.0006126582,0.006828973,0.002336365,0.010201002,0.0022370024,0.0012159866,0.006384517],"category_scores_gemma":[0.039140977,0.00039195883,0.0008461673,0.0067435876,0.0023563907,0.013816828,0.004315493,0.0015342035,0.0017112137],"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":[0.001330871,0.0006094558,0.03253885,0.0013226945,0.00009028323,0.00081315305,0.007705143,0.007713988,0.020418316,0.21907401,0.025268609,0.68311465],"study_design_scores_gemma":[0.00033492697,0.0028414447,0.04879682,0.0036013157,0.00035069228,0.0051329443,0.035576522,0.06540268,0.19924033,0.13284218,0.505121,0.0007592304],"about_ca_topic_score_codex":0.0025577804,"about_ca_topic_score_gemma":0.0024758372,"teacher_disagreement_score":0.017635373,"about_ca_system_score_codex":0.0029869883,"about_ca_system_score_gemma":0.003219687,"threshold_uncertainty_score":0.09326589},"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":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04431867,0.010596381,0.03694451,0.64167297,0.002411833,0.00021710889,0.000044256703,0.0003128032,0.26348153],"genre_scores_gemma":[0.94244736,0.0028004888,0.007835951,0.03343383,0.0017332678,0.0003102449,0.000019941817,0.00020173164,0.011217205],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.8318474,0.1206695,0.004509742,0.007493097,0.024428388,0.011051887],"domain_scores_gemma":[0.669473,0.2558809,0.018475505,0.018287187,0.01676752,0.021115903],"candidate_categories":["scholarly_communication","open_science"],"consensus_categories":[],"category_scores_codex":[0.15418558,0.00071624096,0.0009544038,0.0055525918,0.026709195,0.04764844,0.0043874537,0.01904945,0.011943365],"category_scores_gemma":[0.16000691,0.0011093308,0.0011157581,0.0038636795,0.093731575,0.049799714,0.040821046,0.021640936,0.0018609442],"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":[0.000025611662,0.00003930038,0.00092407264,0.00012340925,0.0000107238575,0.00020031145,0.03463785,0.00013445312,0.00014251115,0.93647856,0.006172891,0.021110347],"study_design_scores_gemma":[0.000056637495,0.00007910724,0.0011226118,0.0013719372,0.000024874826,0.0003876318,0.04421648,0.0010272274,0.0006817448,0.57487845,0.37604797,0.00010545172],"about_ca_topic_score_codex":0.007203439,"about_ca_topic_score_gemma":0.0045241304,"teacher_disagreement_score":0.99561256,"about_ca_system_score_codex":0.022560703,"about_ca_system_score_gemma":0.035344727,"threshold_uncertainty_score":0.815421},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.015477247,0.004152192,0.8433049,0.057708245,0.00023084137,0.0004771695,0.00014215772,0.00055989705,0.077947475],"genre_scores_gemma":[0.3614486,0.0035670442,0.6226159,0.004648307,0.0002962133,0.0009808086,0.0003212205,0.00018742452,0.0059343907],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.951554,0.029337076,0.003575801,0.0028112915,0.010358021,0.0023638909],"domain_scores_gemma":[0.9701802,0.016388543,0.0024697287,0.005520467,0.0037804097,0.0016606358],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.05394041,0.0011863874,0.0021580062,0.0127848005,0.00673748,0.032529544,0.006606008,0.010304273,0.0026674466],"category_scores_gemma":[0.035080984,0.0019168404,0.0027447164,0.011112378,0.030343326,0.04736264,0.019398857,0.011712743,0.00067503593],"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":[0.0000065394233,0.00003843821,0.00033914263,0.00008196156,0.00002513815,0.00007650098,0.002456421,0.002941129,0.00012137613,0.98527294,0.00062589266,0.008014582],"study_design_scores_gemma":[0.000011785077,0.00001571379,0.00027040613,0.0002775731,0.000028139824,0.000093897965,0.004329181,0.0067664427,0.00032639096,0.9636923,0.024162017,0.000026241185],"about_ca_topic_score_codex":0.010635357,"about_ca_topic_score_gemma":0.0063307933,"teacher_disagreement_score":0.96747047,"about_ca_system_score_codex":0.01930285,"about_ca_system_score_gemma":0.028631024,"threshold_uncertainty_score":0.2852676},"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":"methods","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.030615278,0.004603265,0.5209894,0.23121798,0.0018982629,0.0030645262,0.001567192,0.0047983928,0.20124571],"genre_scores_gemma":[0.14498352,0.0034875546,0.77246577,0.011439183,0.0005726234,0.0033035565,0.0046180873,0.00080871675,0.05832107],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9206588,0.039977003,0.0038642555,0.007573981,0.02226641,0.0056595723],"domain_scores_gemma":[0.815643,0.034819666,0.009866955,0.030925544,0.07174882,0.03699602],"candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.1496237,0.00069482974,0.0007205053,0.0066864192,0.012687347,0.021902882,0.008597715,0.005623982,0.008014056],"category_scores_gemma":[0.096513584,0.0011956112,0.0011396295,0.008790148,0.006984667,0.03150968,0.0277707,0.0061032516,0.0035363366],"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":[0.00014004124,0.00033845525,0.012213707,0.00049071276,0.00006354524,0.00044653984,0.017799797,0.0042535583,0.0011471034,0.57512575,0.11745553,0.27052522],"study_design_scores_gemma":[0.000057068,0.00007263319,0.0024220166,0.00061735685,0.000026775691,0.00018520158,0.010834516,0.005610823,0.0007928818,0.08718271,0.892102,0.000096030744],"about_ca_topic_score_codex":0.10321951,"about_ca_topic_score_gemma":0.095944,"teacher_disagreement_score":0.99140227,"about_ca_system_score_codex":0.02988997,"about_ca_system_score_gemma":0.14980556,"threshold_uncertainty_score":0.79129523},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.55477947,0.2213556,0.097295344,0.032450113,0.00092129194,0.053541694,0.01599824,0.00028236266,0.023375835],"genre_scores_gemma":[0.6713642,0.11380945,0.13317569,0.007241928,0.00023898704,0.0652506,0.0061552944,0.0001876994,0.0025762795],"study_design_codex":"qualitative","study_design_gemma":"systematic_review","domain_scores_codex":[0.87374943,0.07712789,0.026014552,0.0049134456,0.015530417,0.0026642284],"domain_scores_gemma":[0.5115516,0.42762876,0.019129181,0.007669001,0.032018702,0.0020027366],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.16940682,0.0009641079,0.0028963636,0.024231562,0.0035438873,0.006029914,0.0025676128,0.0028693432,0.0027502757],"category_scores_gemma":[0.24989907,0.0010987814,0.0018261021,0.033985678,0.0041930275,0.011466835,0.0058258367,0.0016755938,0.00060308684],"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":[0.00043501562,0.00043747024,0.024633024,0.28872123,0.0006986316,0.0038925777,0.3930746,0.0007431116,0.008438626,0.008895011,0.006664918,0.26336583],"study_design_scores_gemma":[0.00021332326,0.00066185254,0.021390855,0.26831606,0.0024560464,0.0018068539,0.5146836,0.000943987,0.007235236,0.00991383,0.1721728,0.00020557469],"about_ca_topic_score_codex":0.0049915803,"about_ca_topic_score_gemma":0.018279653,"teacher_disagreement_score":0.9939701,"about_ca_system_score_codex":0.007822537,"about_ca_system_score_gemma":0.03957597,"threshold_uncertainty_score":0.89591956},"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":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.44793895,0.023404326,0.0017765348,0.38270015,0.002515848,0.0006053389,0.00024476604,0.00016480313,0.14064936],"genre_scores_gemma":[0.8710933,0.017066924,0.0070876665,0.028199105,0.0003093953,0.00078096235,0.0002588346,0.000078120334,0.07512572],"study_design_codex":"qualitative","study_design_gemma":"not_applicable","domain_scores_codex":[0.99624264,0.0019007057,0.00014755277,0.00024844325,0.0003597364,0.001100976],"domain_scores_gemma":[0.9936395,0.0010208496,0.00033704747,0.0001613402,0.00035737376,0.004483906],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0071721,0.00032385768,0.00040520742,0.001147928,0.011917547,0.004647875,0.0014256096,0.0032310283,0.015173763],"category_scores_gemma":[0.008434664,0.0005188959,0.00029979035,0.0014285763,0.0034566068,0.0071295453,0.013693193,0.0035220524,0.001987571],"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":[0.00012368572,0.0012594275,0.018637272,0.0021902462,0.000016540223,0.0041435855,0.54810995,0.0002139452,0.0031333475,0.032061346,0.090955555,0.29915506],"study_design_scores_gemma":[0.000032624783,0.0002453627,0.017230943,0.0026771245,0.000008464335,0.00084698375,0.3254011,0.00015678829,0.0006299171,0.0039846813,0.64874715,0.000038950126],"about_ca_topic_score_codex":0.0142797865,"about_ca_topic_score_gemma":0.035236098,"teacher_disagreement_score":0.99535215,"about_ca_system_score_codex":0.0071448553,"about_ca_system_score_gemma":0.03656303,"threshold_uncertainty_score":0.05183977},"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":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.88487333,0.0084624225,0.006995596,0.032143608,0.00050812605,0.0003999722,0.0028536296,0.0008244649,0.06293884],"genre_scores_gemma":[0.9790851,0.00308741,0.0052629267,0.002148379,0.0002573763,0.0002003386,0.00084162265,0.00029236515,0.008824421],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.940758,0.024936317,0.010395132,0.0071823164,0.012033658,0.0046946313],"domain_scores_gemma":[0.6873927,0.12830979,0.076238215,0.033272177,0.053234145,0.021552984],"candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.08020264,0.00027392243,0.0007367121,0.009741816,0.003277003,0.014027988,0.0033284256,0.0013773633,0.00892552],"category_scores_gemma":[0.21359576,0.00081419246,0.0007482907,0.017760165,0.004546058,0.008615649,0.008578263,0.0015454743,0.0024811649],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011766003,0.00048274084,0.49362385,0.00474554,0.00048880035,0.0016578789,0.06654738,0.004362193,0.0076723173,0.047157012,0.027367724,0.3447181],"study_design_scores_gemma":[0.00015279099,0.00060252025,0.5335543,0.0026097437,0.00020725743,0.0014359829,0.09771829,0.0033718257,0.01056114,0.010913611,0.33838108,0.0004914003],"about_ca_topic_score_codex":0.02045057,"about_ca_topic_score_gemma":0.019112421,"teacher_disagreement_score":0.99667156,"about_ca_system_score_codex":0.011168348,"about_ca_system_score_gemma":0.017860914,"threshold_uncertainty_score":0.42415714},"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":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.18387803,0.0006268335,0.74361813,0.006974016,0.00068663224,0.02206369,0.0010203752,0.0038868105,0.037245486],"genre_scores_gemma":[0.1869724,0.00038583908,0.7898956,0.0006736482,0.0000393104,0.015345135,0.00050922285,0.00040558152,0.005773212],"study_design_codex":"design_other","study_design_gemma":"qualitative","domain_scores_codex":[0.9416327,0.046974573,0.0021976933,0.003020274,0.004588761,0.0015860754],"domain_scores_gemma":[0.9194216,0.053778317,0.0020340434,0.011909459,0.009997563,0.0028590008],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.07438054,0.000991734,0.00061380933,0.0022495666,0.0057808496,0.008022252,0.004909436,0.0024270918,0.00689077],"category_scores_gemma":[0.09322382,0.0011356393,0.0011479231,0.0015475206,0.006250484,0.008659081,0.009732671,0.0021396198,0.0017046559],"study_design_candidate":"qualitative","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":[0.0009950845,0.0030171883,0.01183565,0.005275438,0.0001358767,0.0025362512,0.38996378,0.00636022,0.024299674,0.074385464,0.024414277,0.45678112],"study_design_scores_gemma":[0.0016115482,0.005930209,0.007986805,0.00377354,0.00029609434,0.0017168598,0.23772319,0.020870902,0.04219601,0.073357865,0.60392684,0.0006102405],"about_ca_topic_score_codex":0.002826774,"about_ca_topic_score_gemma":0.004269335,"teacher_disagreement_score":0.9256195,"about_ca_system_score_codex":0.0034012448,"about_ca_system_score_gemma":0.0144294165,"threshold_uncertainty_score":0.39336658},"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":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9400627,0.001053916,0.046925914,0.0014777543,0.000039338905,0.00017918982,0.0014907087,0.0004452784,0.00832517],"genre_scores_gemma":[0.95856845,0.00037601186,0.03826019,0.00006749066,0.00004093774,0.00018363721,0.0010132568,0.000107847314,0.0013822233],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9944166,0.002754002,0.00042084258,0.00087636465,0.0010401848,0.0004920099],"domain_scores_gemma":[0.95754266,0.030835317,0.004676598,0.0030079884,0.0028735015,0.0010638904],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0073199086,0.00034080195,0.00046142266,0.013544821,0.00221943,0.0033532572,0.000669614,0.0010413408,0.0017623033],"category_scores_gemma":[0.025044415,0.00034795015,0.0006600866,0.016192522,0.001243965,0.003589553,0.002418752,0.00093439437,0.0003221326],"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":[0.0004270941,0.0004294569,0.46188858,0.0010750978,0.00011532724,0.0034651768,0.20044728,0.005493752,0.0139327785,0.026303913,0.00715579,0.2792657],"study_design_scores_gemma":[0.00006762395,0.00025961036,0.58112055,0.0007597129,0.00023932375,0.0041400664,0.16274552,0.10816875,0.013866956,0.025071146,0.10333484,0.00022586422],"about_ca_topic_score_codex":0.009703952,"about_ca_topic_score_gemma":0.011918996,"teacher_disagreement_score":0.99664676,"about_ca_system_score_codex":0.0017892579,"about_ca_system_score_gemma":0.0017389345,"threshold_uncertainty_score":0.038711846},"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":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.033289958,0.011114122,0.46364045,0.2299985,0.004957349,0.00038736072,0.0003629587,0.00092098286,0.25532827],"genre_scores_gemma":[0.5636455,0.018083647,0.3122017,0.024027247,0.0019465814,0.0011284525,0.0007423708,0.0010612392,0.077163175],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","domain_scores_codex":[0.97404987,0.015158993,0.0010607205,0.0027566627,0.005932054,0.0010417216],"domain_scores_gemma":[0.9604982,0.017488983,0.0051970254,0.007445577,0.0076148934,0.0017553307],"candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.030158693,0.00068305206,0.0008391772,0.0069385557,0.009929478,0.02767865,0.0026925695,0.005447825,0.009284561],"category_scores_gemma":[0.037803452,0.00067366,0.00089999073,0.008647525,0.03137359,0.03569191,0.013340699,0.0070754336,0.0032913052],"study_design_candidate":"qualitative","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":[0.00003225538,0.0000435034,0.0014911921,0.00034439127,0.000017437253,0.00011949384,0.03364011,0.00036927342,0.00065472134,0.8877506,0.020373873,0.05516319],"study_design_scores_gemma":[0.00001228053,0.000040578605,0.0010519584,0.000931768,0.000018483779,0.00035304372,0.039793514,0.00089076604,0.00089813693,0.37414882,0.5817914,0.00006936486],"about_ca_topic_score_codex":0.0062852986,"about_ca_topic_score_gemma":0.006636462,"teacher_disagreement_score":0.9973074,"about_ca_system_score_codex":0.006245057,"about_ca_system_score_gemma":0.011229763,"threshold_uncertainty_score":0.15949637},"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":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.078412905,0.00090838195,0.84331053,0.021357475,0.00088504556,0.0022447887,0.0042289393,0.004701475,0.043950506],"genre_scores_gemma":[0.24900123,0.0005585157,0.72852457,0.0035935778,0.00037497972,0.0022502118,0.0049531064,0.001809181,0.0089345975],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.925289,0.03534051,0.014340144,0.007448556,0.01407301,0.003508773],"domain_scores_gemma":[0.78830737,0.07939839,0.007837107,0.07251128,0.045594715,0.0063510695],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.083822764,0.0005042372,0.001375238,0.012439621,0.009239274,0.018996814,0.0044332556,0.0026583907,0.0058716587],"category_scores_gemma":[0.14407092,0.0010500042,0.0018237812,0.009805987,0.015149293,0.03257061,0.015606853,0.007199634,0.0017079593],"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":[0.00017643918,0.00010168554,0.0072241514,0.0004330495,0.000044903492,0.00025190823,0.02376173,0.001727119,0.0025033578,0.84412754,0.015369323,0.10427875],"study_design_scores_gemma":[0.000066251225,0.00014998444,0.0061225523,0.001087129,0.000081136706,0.00045962125,0.022133904,0.008014148,0.0058640866,0.35357434,0.60218346,0.0002633778],"about_ca_topic_score_codex":0.030134972,"about_ca_topic_score_gemma":0.021466624,"teacher_disagreement_score":0.98100317,"about_ca_system_score_codex":0.009293781,"about_ca_system_score_gemma":0.017364051,"threshold_uncertainty_score":0.44330245},"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":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.85493654,0.009141172,0.010551208,0.09683007,0.00040953848,0.0006733522,0.00017861449,0.00029579038,0.026983758],"genre_scores_gemma":[0.9762564,0.0039856485,0.012932936,0.0039368966,0.00007403643,0.00030460837,0.000072847455,0.000040923856,0.0023957088],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.95535725,0.026983313,0.003475996,0.0026183315,0.008338214,0.0032268665],"domain_scores_gemma":[0.90529406,0.050029002,0.01262343,0.008400224,0.013366432,0.010286803],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.050720356,0.0003512624,0.00058489153,0.003018885,0.007911606,0.017788786,0.0029465926,0.0029828432,0.0028274243],"category_scores_gemma":[0.10301575,0.0007993532,0.0006161618,0.0055564553,0.0067747864,0.011854875,0.010099421,0.0041378746,0.0005669276],"study_design_candidate":"observational","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":[0.00026049672,0.0004501551,0.11872853,0.002209988,0.00007395134,0.0027209565,0.38381568,0.00041212025,0.00871254,0.022459326,0.0043065334,0.45584968],"study_design_scores_gemma":[0.000046140183,0.00059628265,0.14263317,0.004008706,0.00005387075,0.0015515737,0.6382503,0.0012045056,0.0051563126,0.006795531,0.199461,0.00024253041],"about_ca_topic_score_codex":0.022818182,"about_ca_topic_score_gemma":0.019745372,"teacher_disagreement_score":0.98221123,"about_ca_system_score_codex":0.018537791,"about_ca_system_score_gemma":0.042550128,"threshold_uncertainty_score":0.26823807},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.28456533,0.0035163548,0.4856561,0.0034409084,0.0015683457,0.0015694229,0.0047254437,0.114729635,0.10022837],"genre_scores_gemma":[0.79409915,0.0012429714,0.15387066,0.0009952603,0.00018366956,0.00039275357,0.0046942728,0.003940797,0.04058051],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99893135,0.000073831536,0.000063745174,0.00012310967,0.0005877122,0.00022026392],"domain_scores_gemma":[0.998288,0.0001743765,0.000120822086,0.0005220981,0.000606752,0.00028789634],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010615739,0.00039955578,0.00042624198,0.00084813253,0.0015433127,0.0028067892,0.0024008867,0.0012398314,0.0072962577],"category_scores_gemma":[0.0027785986,0.00029701344,0.00038020447,0.0012512739,0.0014605083,0.004651611,0.003015706,0.0009126784,0.003466501],"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":[0.004839742,0.00056034286,0.010789306,0.0019965442,0.0002004648,0.003617958,0.0043171565,0.015334591,0.16693649,0.15282314,0.21302348,0.4255608],"study_design_scores_gemma":[0.00030510713,0.0007702653,0.008195692,0.00041563992,0.00018225434,0.003874144,0.001561394,0.10284096,0.1941538,0.051344387,0.6359104,0.0004460284],"about_ca_topic_score_codex":0.005798446,"about_ca_topic_score_gemma":0.004952702,"teacher_disagreement_score":0.0072962577,"about_ca_system_score_codex":0.0015252199,"about_ca_system_score_gemma":0.0023093477,"threshold_uncertainty_score":0.02440846},"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":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.017438417,0.0007568238,0.9599884,0.0042834855,0.0000644489,0.00059152464,0.0013116717,0.0019380363,0.013627175],"genre_scores_gemma":[0.19511037,0.001021553,0.79428893,0.00047816982,0.00006879315,0.0007883452,0.003855264,0.0004341861,0.0039544124],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.982928,0.005135658,0.00299552,0.0021499717,0.0057992884,0.0009915336],"domain_scores_gemma":[0.94434124,0.020838825,0.0050009456,0.007562802,0.020177975,0.002078222],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.028517082,0.00081159873,0.0006909743,0.010577598,0.002313699,0.01020194,0.0027783914,0.0026092762,0.0025282863],"category_scores_gemma":[0.06403367,0.0011665424,0.0018511553,0.0074308305,0.0023575511,0.0210342,0.006388663,0.0033672475,0.0016577803],"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":[0.00013786304,0.0001681555,0.01403142,0.00062073854,0.000078739504,0.0003941592,0.0041475394,0.02038023,0.0035959438,0.8237921,0.008001392,0.12465174],"study_design_scores_gemma":[0.000050798306,0.00026940482,0.0054978495,0.0012713464,0.00013430354,0.0011714291,0.00305983,0.20907888,0.007819428,0.6039662,0.16743112,0.0002494773],"about_ca_topic_score_codex":0.011070807,"about_ca_topic_score_gemma":0.009774327,"teacher_disagreement_score":0.98979807,"about_ca_system_score_codex":0.007207024,"about_ca_system_score_gemma":0.008259223,"threshold_uncertainty_score":0.15081453},"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":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6104767,0.0073436396,0.22345574,0.061255507,0.0008040397,0.012221846,0.0070364466,0.0007791088,0.076627016],"genre_scores_gemma":[0.85935444,0.0014561097,0.116845936,0.0050141565,0.00031670774,0.0073400834,0.001914598,0.00028100447,0.007476916],"study_design_codex":"qualitative","study_design_gemma":"qualitative","domain_scores_codex":[0.45607516,0.37371147,0.07546894,0.031091774,0.055949118,0.00770357],"domain_scores_gemma":[0.2745044,0.47809935,0.0684485,0.12100297,0.046601687,0.011343065],"candidate_categories":["metaresearch","open_science"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.34574565,0.00056435645,0.0016892103,0.015235768,0.020937037,0.019695103,0.0056275376,0.0052770595,0.006732424],"category_scores_gemma":[0.44842118,0.0018603436,0.0011266008,0.02550348,0.028291048,0.021405406,0.02425319,0.0049993563,0.0015106603],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00038694285,0.00019348072,0.054297905,0.0019011622,0.00015084253,0.0039010504,0.6635988,0.00092254684,0.0023227222,0.17779581,0.008673819,0.08585495],"study_design_scores_gemma":[0.00017334033,0.0003259961,0.039908472,0.0077750944,0.00015378052,0.0038600059,0.42786786,0.0025982794,0.0037165165,0.11246863,0.4005612,0.0005907552],"about_ca_topic_score_codex":0.025816403,"about_ca_topic_score_gemma":0.036592033,"teacher_disagreement_score":0.9943725,"about_ca_system_score_codex":0.024969075,"about_ca_system_score_gemma":0.062029034,"threshold_uncertainty_score":0.8068121},"labels":[],"label_agreement":null}]}