2026-10-11T17:56:21Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33975403 2026-09-23T17:28:51Z category_146 category_7 category_14 category_272 category_45 category_734 category_931 category_64 portal_5 item_type_3 month_year_09_2026

<p>Predictive performance of candidate models in MIMIC IV Dataset.</p> Weisheng Chen (548014) Xinya Li (477111) Zhongkai Qu (25099054) Ting Liu (45625) Bo Chen (32294) Liying Huang (366417) Ningxia He (12864170) Zhigang Wang (13556) Jun Lyu (6201101) Space Science Medicine Molecular Biology Environmental Sciences not elsewhere classified Sociology Biological Sciences not elsewhere classified Information Systems not elsewhere classified Cancer subsequently performed using structured codes rather related case identification point estimates calculated increased computational costs identification relies heavily critically ill patients analyzed using pre proportional performance gains evaluated using accuracy interpreting radiology reports detecting radiographic evidence xlink "> pneumonia large language models radiology reports radiographic evidence trained models radiological reports radiological evidence model performance tertiary hospital radiological text poor outcomes parameter size parameter scale methodological basis may limit knowledge augmentation iv dataset iv database independent evaluation increase linearly highest accuracy gz dataset full dataset enhanced robustness dataset splitting corresponding 95 consistently translate common infection automated detection augmented generation 27b achieving 000 patient <p>Predictive performance of candidate models in MIMIC IV Dataset.</p> 2026-09-23T17:28:41Z Dataset Dataset 10.1371/journal.pdig.0001104.t003 https://figshare.com/articles/dataset/_p_Predictive_performance_of_candidate_models_in_MIMIC_IV_Dataset_p_/33975403 CC BY 4.0