{"id":"W4376869324","doi":"10.18280/isi.280207","title":"A Schema Integration Approach for Big Data Analysis","year":2023,"lang":"en","type":"article","venue":"Ingénierie des systèmes d information","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Schema matching; NoSQL; Computer science; Data integration; Star schema; Schema (genetic algorithms); Conceptual schema; Database schema; Information schema; Schema evolution; Data mining; Schema migration; Information retrieval; Data science; Big data; Semi-structured model; Database design","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007365352,0.001097505,0.001013974,0.005808887,0.001333906,0.005024896,0.002584007,0.001572753,0.003535643],"category_scores_gemma":[0.01050328,0.0009421549,0.003185709,0.00748526,0.0009832582,0.006526474,0.004605039,0.002576435,0.00151791],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001145825,"about_ca_system_score_gemma":0.002830256,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003822097,"about_ca_topic_score_gemma":0.0052978,"domain_scores_codex":[0.9932142,0.001968661,0.001006138,0.00102747,0.00257706,0.0002064041],"domain_scores_gemma":[0.9946626,0.001283725,0.0004267054,0.002003697,0.001402837,0.0002203092],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003541367,0.0005910376,0.01011721,0.001405827,0.00102244,0.001369635,0.003027493,0.03617326,0.02092331,0.2899778,0.02769161,0.6073462],"study_design_scores_gemma":[0.00007537718,0.0001781501,0.003651906,0.0005546567,0.000426125,0.00185719,0.001827346,0.4754257,0.02623985,0.2602323,0.2293834,0.0001480604],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001713933,0.0004414451,0.9928548,0.0002620747,0.00005729435,0.0002340537,0.0004988299,0.002379077,0.001558569],"genre_scores_gemma":[0.0225327,0.0004744805,0.9717264,0.0002293692,0.00003184741,0.0002410248,0.003311381,0.0003527854,0.001100011],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007365352,"threshold_uncertainty_score":0.03895211,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.292872492286068,"score_gpt":0.3912052498639068,"score_spread":0.09833275757783871,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}