2026-10-14T02:36:16Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34016436 2026-09-28T17:34:19Z category_15 category_21 category_734 category_931 category_106 portal_5 item_type_3 month_year_09_2026

<p>Class-wise performance metrics (macro-averaged, one-vs-rest) of the TPE-Optimized XGBoost model on the held-out test set.</p> Juntao Chen (6415844) Jinmei Zhan (25132656) Xiaodeng Zhou (20300145) Jiahua Yao (20300148) Su-Kit Tang (20300151) Xia Liu (117318) Neuroscience Biotechnology Biological Sciences not elsewhere classified Information Systems not elsewhere classified Science Policy volunteer service hours tertiary vocational colleges structured parzen estimator proposed model achieves mitigate overfitting risk dataset comprises 2 8 %, 3 8 %, 17 22 informative variables 2016 &# 8211 rf ), k recursive feature elimination xgboost classifier using tpe hyperparameter optimization fine &# 8211 optimized xgboost model optimized xgboost knn ), feature set urgent need three fine test split test set stratified 5 smooth employment overall accuracy nearest neighbor logistic regression including non grained prediction grained classes gaokao scores fold cross extracurricular engagement entrance scores employment destinations class distribution balanced accuracy averaged specificity averaged recall averaged precision averaged f1 academic records 4 %). 270 graduates 2019 cohorts 20 train <p>Class-wise performance metrics (macro-averaged, one-vs-rest) of the TPE-Optimized XGBoost model on the held-out test set.</p> 2026-09-28T17:33:58Z Dataset Dataset 10.1371/journal.pone.0359421.t006 https://figshare.com/articles/dataset/_p_Class-wise_performance_metrics_macro-averaged_one-vs-rest_of_the_TPE-Optimized_XGBoost_model_on_the_held-out_test_set_p_/34016436 CC BY 4.0