Rui Yu

Mr. Rui Yu

orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 September 2026 in Journal of King Saud University Computer and Information Sciences
Verified
Journal Article
Open access
PDF file
HZHuanze Zeng
Huanze Zeng
RXRui Xu
Rui Xu
ALAoqi Lu
Aoqi Lu
RYRui Yu
Rui Yu
AZAnhao Zhang
Anhao Zhang
TLTing Li
Ting Li
In the intensive care unit (ICU), the incidence of acute kidney injury (AKI) can exceed 50%, and due to its complex and difficult treatment process, the mortality rate of AKI is significantly correlated with the incidence rate. Therefore, to save patients’ lives and improve the prognosis, the early diagnosis and prediction of AKI are crucial. Most existing AKI studies based on the medical information mart for intensive care III (MIMIC-III) dataset only use patients’ static data for prediction. This study proposes a novel interpretable dynamic AKI prediction method based on the ClinicalBERT-temporal fusion transformer (TFT). This method uses TFT to extract features from structured time-series clinical data and ClinicalBERT to extract features from unstructured clinical text. Then the fused multimodal features are input into a multilayer perceptron to predict AKI. Additionally, the proposed method can output the ranking results of feature importance. While enhancing the model’s credibility and transparency, it enables personalized AKI prediction and offers multi-dimensional data support for developing treatment and prognosis plans. In the comparative experiments with common AKI prediction models, the proposed method demonstrated higher performance, proving feasibility and advancement of the proposed multimodal fusion AKI prediction based on time series dynamic data.