{"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":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.007989397,0.0001312492,0.0002850836,0.001413762,0.0003172528,0.001163621,0.00131151,0.00007723133,0.00002837619],"category_scores_gemma":[0.005971925,0.0001027582,0.0001240985,0.004880618,0.00008853571,0.005475838,0.0005456065,0.00006520182,0.0004734217],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007566305,"about_ca_system_score_gemma":0.00005161129,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007560856,"about_ca_topic_score_gemma":0.00007084526,"domain_scores_codex":[0.9973921,0.000111648,0.001006877,0.0003226843,0.0009088598,0.0002578977],"domain_scores_gemma":[0.997207,0.0004242371,0.0004380381,0.001456148,0.0004074084,0.00006713851],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00004724116,0.00003060704,0.0005198239,0.0001136909,0.0002672204,4.01707e-7,0.003941567,0.003166655,0.00002413466,0.01924262,0.09353653,0.8791095],"study_design_scores_gemma":[0.0004038144,0.00004912189,0.006166896,0.00001819528,0.0002193651,0.000001633074,0.01634694,0.7699723,0.0001500407,0.03039641,0.176008,0.000267211],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006795284,0.00001246857,0.9808017,0.0001859347,0.000395287,0.0005057173,0.0007035131,0.0001977835,0.01040233],"genre_scores_gemma":[0.9492772,0.00002307843,0.02656776,0.0004286687,0.0002166147,0.0002193226,0.02205095,0.000009537615,0.001206905],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9542339,"threshold_uncertainty_score":0.9998733,"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."}}