{"id":"W1996324618","doi":"10.1109/cse.2013.129","title":"Topics and Terms Mining in Unstructured Data Stores","year":2013,"lang":"en","type":"article","venue":"","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"NoSQL; Computer science; Unstructured data; Data mining; Schema (genetic algorithms); Database; Big data; Knowledge extraction; Semi-structured data; Data extraction; Data warehouse; Information retrieval; Relational database","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002736424,0.0008939813,0.001285478,0.005756547,0.001423636,0.004753646,0.00171257,0.0007720227,0.002162532],"category_scores_gemma":[0.0160114,0.000651543,0.001342048,0.007896693,0.0007921858,0.006767882,0.003134629,0.0009698102,0.00228724],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009043252,"about_ca_system_score_gemma":0.002007056,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004781882,"about_ca_topic_score_gemma":0.006685368,"domain_scores_codex":[0.9964789,0.0005456461,0.0005013563,0.0007196002,0.001511129,0.0002433287],"domain_scores_gemma":[0.9926559,0.002867423,0.0006018806,0.002194701,0.001449976,0.0002302893],"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":[0.00133858,0.0003584114,0.0265182,0.001550688,0.0002367365,0.001263649,0.003826328,0.03464504,0.03637758,0.0714937,0.04756452,0.7748266],"study_design_scores_gemma":[0.0002081289,0.000229036,0.0111053,0.0002872499,0.0001549484,0.001500707,0.004339011,0.5074081,0.08947921,0.2536166,0.1315321,0.0001396983],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1010468,0.001779325,0.8370946,0.001476536,0.0002208879,0.001037928,0.01939309,0.02676301,0.01118787],"genre_scores_gemma":[0.1740271,0.0009089041,0.7996187,0.0002054869,0.00008674067,0.0004134385,0.01978156,0.0007853137,0.004172739],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005756547,"threshold_uncertainty_score":0.01447177,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02813821789009692,"score_gpt":0.2605091579314912,"score_spread":0.2323709400413943,"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."}}