scholarly journals Using Features Extracted from Vital Time Series for Early Prediction of Sepsis

2019 ◽  
Author(s):  
Qiang Yu ◽  
Xiaolin Huang ◽  
Weifeng Li ◽  
Cheng Wang ◽  
Ying Chen ◽  
...  
Keyword(s):  
Time Series ◽  
Early Prediction

scholarly journals Multiscale network representation of physiological time series for early prediction of sepsis

2017 ◽  
Vol 38 (12) ◽  
pp. 2235-2248 ◽  
Author(s):  
Supreeth P Shashikumar ◽  
Qiao Li ◽  
Gari D Clifford ◽  
Shamim Nemati
Keyword(s):  
Time Series ◽  
Early Prediction ◽  
Network Representation ◽  
Physiological Time

scholarly journals Classification of cardiovascular time series based on different coupling structures using recurrence networks analysis

2013 ◽  
Vol 371 (1997) ◽  
pp. 20110623 ◽  
Author(s):  
Gonzalo Marcelo Ramírez Ávila ◽  
Andrej Gapelyuk ◽  
Norbert Marwan ◽  
Thomas Walther ◽  
Holger Stepan ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Heart Rate ◽  
Time Series ◽  
Path Length ◽  
Clustering Coefficient ◽  
Early Prediction ◽  
Quadratic Discriminant Analysis ◽  
Novel Approach ◽  
Networks Analysis

We analyse cardiovascular time series with the aim of performing early prediction of preeclampsia (PE), a pregnancy-specific disorder causing maternal and foetal morbidity and mortality. The analysis is made using a novel approach, namely the ε -recurrence networks applied to a phase space constructed by means of the time series of the variabilities of the heart rate and the blood pressure (systolic and diastolic). All the possible coupling structures among these variables are considered for the analysis. Network measures such as average path length, mean coreness, global clustering coefficient and scale-local transitivity dimension are computed and constitute the parameters for the subsequent quadratic discriminant analysis. This allows us to predict PE with a sensitivity of 91.7 per cent and a specificity of 68.1 per cent, thus validating the use of this method for classifying healthy and preeclamptic patients.

Early Prediction of Lean Blowout from Chemiluminescence Time Series Data

2020 ◽  
pp. 1-28
Author(s):  
Sudeepta Mondal ◽  
Somnath De ◽  
Achintya Mukhopadhyay ◽  
Swarnendu Sen ◽  
Asok Ray
Keyword(s):  
Time Series ◽  
Time Series Data ◽  
Series Data ◽  
Early Prediction ◽  
Lean Blowout

scholarly journals Utilizing imbalanced electronic health records to predict acute kidney injury by ensemble learning and time series model

2020 ◽  
Vol 20 (1) ◽  
Author(s):  
Yuan Wang ◽  
Yake Wei ◽  
Hao Yang ◽  
Jingwei Li ◽  
Yubo Zhou ◽  
...  
Keyword(s):  
Time Series ◽  
Acute Kidney Injury ◽  
Ensemble Learning ◽  
Class Size ◽  
Model Performance ◽  
Kidney Injury ◽  
Time Series Model ◽  
Early Prediction ◽  
Medication Information ◽  
Mimic Iii

Abstract Background Acute Kidney Injury (AKI) is a shared complication among Intensive Care Unit (ICU), marked by high cost, high morbidity and high mortality. As the early prediction of AKI is critical for patients’ outcomes and data mining is such a powerful prediction tool, many AKI prediction models based on machine learning methods have been proposed. Our motivation is inspired by the fact that the incidence of AKI is a changing temporal sequence affected by the joint action of patients’ daily drug combinations and their physiological indexes. However, most existing models have not considered such a temporal correlation. Besides, due to great challenges caused by sparse, high-dimensional and highly imbalanced clinical data, it is hard to achieve ideal performance. Methods We develop a fast, simple and less-costly model based on an ensemble learning algorithm, named Ensemble Time Series Model (ETSM). Besides benefiting from vital signs and laboratory results as explicit indicators, ETSM explores the effect of drug combinations as possible implicit indicators for the AKI prediction. The model transforms temporal medication information into a multidimensional vector to consider and measure drug cumulative effects that may cause AKI. Results We compare ETSM with state-of-the-art models on ICUC and MIMIC III datasets. On the basis of the experimental results, our model obtains satisfactory performance (ICUC: AUC 24 hours ahead: 0.81, 48 hours ahead: 0.78; MIMIC III: AUC 24 hours ahead: 0.95, 48 hours ahead: 0.95). Meanwhile, we compare the effects of different sampling and feature generation methods on the model performance. In the ablation study, we validate that medication information improves model performance (24 hours ahead: AUC increased from 0.74 to 0.81). We also find that the model’s performance is closely related to the balanced level of the derivation dataset. The optimal ratio of major class size to minor class size for the model is found for AKI prediction. Conclusions ETSM is an effective method for the early prediction of AKI. The model verifies that AKI incidence is related to the clinical medication. In comparison with other prediction methods, ETSM provides comparable performance results and better interpretability.

scholarly journals An Interpretable Deep Learning Model for Early Prediction of Sepsis in the Emergency Department

2020 ◽  
Author(s):  
Dongdong Zhang ◽  
Changchang Yin ◽  
Katherine M. Hunold ◽  
Xiaoqian Jiang ◽  
Jeffrey M. Caterino ◽  
...  
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Data Science ◽  
Time Series Prediction ◽  
Learning Model ◽  
Attention Mechanism ◽  
Early Prediction ◽  
Max Pooling ◽  
Time Encoding ◽  
Deep Learning Model

Background: Sepsis, a life-threatening illness caused by the body's response to an infection, is the leading cause of death worldwide and has become a global epidemiological burden. Early prediction of sepsis increases the likelihood of survival for septic patients. Methods The 2019 DII National Data Science Challenge enabled participating teams to develop models for early prediction of sepsis onset with de-identified electronic health records of over 100,000 unique patients. Our task is to predict sepsis onset 4 hours before its diagnosis using basic administrative and demographics, time-series vital, lab, nutrition as features. An LSTM-based model with event embedding and time encoding is proposed to model time-series prediction. We utilized the attention mechanism and global max pooling techniques to enable interpretation for the proposed deep learning model. Results We evaluated the performance of the proposed model on 2 use cases of sepsis onset prediction which achieved AUC scores of 0.940 and 0.845, respectively. Our team, BuckeyeAI achieved an average AUC of 0.892 and the official rank is #2 out of 30 participants. Conclusions Our model outperformed collapsed models (i.e., logistic regression, random forest, and LightGBM). The proposed LSTM-based model handles irregular time intervals by incorporating time encoding and is interpretable thanks to the attention mechanism and global max pooling techniques.

Sign in / Sign up

Export Citation Format

Share Document