http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
도탄공 ( Thanh-cong Do ),양형정 ( Hyung-jeong Yang ),김수형 ( Soo-hyung Kim ),고보건 ( Bo-gun Kho ) 사단법인 한국빅데이터서비스학회 2023 빅데이터서비스학회 논문집 Vol.1 No.1
In hospitals globally, the occurrence of clinical deterioration within the hospital setting poses a significant burden on healthcare. Rapid clinical intervention becomes a crucial task in such cases. To detect early signs of clinical decline and prevent cardiac arrest, the intensive care units (ICUs) widely employ rapid response systems (RRSs). In recent years, there has been an increasing use of deep learning (DL) and electronic health records (EHR) in the intensive care domain, such as the prediction of cardiac arrest, sepsis, or transferring to ICUs. DL-based applications are able to learn from sequential time-series data. However, their lack of interpretability leads to low sensitivity and high late alarm rates, and are thus difficult to apply to the clinical settings. In this research, we propose an interpretable end-to-end deep learning architecture that interpolates high-dimensional sequential data and detects the abnormal status of cardiac arrest patients. The experiments are conducted on a private clinical dataset collected from Chonnam National University hospital (CNUH). The experimental results have shown the potential performance of our model, compared to some other state-of-the-art methods.