2026-10-11T10:24:42Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34016669 2026-09-28T17:36:20Z category_12 category_734 category_811 category_64 portal_5 item_type_3 month_year_09_2026

<p>Performance of ML models for classification of pre-MetS and MetS.</p> Metha Yaikwawong (23672778) Khanittha Kamdee (23672781) Somlak Chuengsamarn (23672787) Cell Biology Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Cancer xlink "> among ten evaluated algorithms performed exclusively within fasting plasma glucose ensemble methods achieved 970 &# 8211 906 &# 8211 905 &# 8211 885 &# 8211 026 &# 8211 subsequently evaluated using center retrospective cross 080 ), followed routine clinical variables fixed threshold criteria although substantial discrimination related derived indices lowest brier score least one idf identified features contributing 657 thai adults distinguishing metabolic states test dataset (< metabolic classification tasks mets diagnostic components highest classification performance mets diagnostic criteria thai adults features contributing lowest discrimination highest performance highest discrimination diagnostic definition trained using score normalization fold cross 969 ), metabolic syndrome metabolic dysfunction training dataset observed performance four clinical clinical implementation xgboost classification xgboost demonstrated xgboost achieving <p>Performance of ML models for classification of pre-MetS and MetS.</p> 2026-09-28T17:36:01Z Dataset Dataset 10.1371/journal.pone.0359283.t002 https://figshare.com/articles/dataset/_p_Performance_of_ML_models_for_classification_of_pre-MetS_and_MetS_p_/34016669 CC BY 4.0