ICU Survival Prediction Incorporating Test-Time Augmentation to Improve the Accuracy of Ensemble-Based Models

Cohen, Seffi; Dagan, Noa; Cohen-Inger, Nurit; Ofer, Dan; Rokach, Lior

doi:10.1109/access.2021.3091622

Cited by 25 publications

(20 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Feature extraction, engineering, selection ( Ofer et al, 2021 ), and ML model selection, parameter tuning, and training were performed using the SparkBeyond autoML framework (See patent/US20170017900A1). Previous work has shown the benefit autoML models, in order to comprehensively and automatically find possible predictive signals in complex data, including in biology and healthcare ( Cohen et al, 2021 ). The system automatically extracts and ranks a wide range of compositional features from training data.…”

Section: Methodsmentioning

confidence: 99%

Inferring microRNA regulation: A proteome perspective

Ofer

Linial

2022

Front. Mol. Biosci.

Self Cite

View full text Add to dashboard Cite

Post-transcriptional regulation in multicellular organisms is mediated by microRNAs. However, the principles that determine if a gene is regulated by miRNAs are poorly understood. Previous works focused mostly on miRNA seed matches and other features of the 3′-UTR of transcripts. These common approaches rely on knowledge of the miRNA families, and computational approaches still yield poor, inconsistent results, with many false positives. In this work, we present a different paradigm for predicting miRNA-regulated genes based on the encoded proteins. In a novel, automated machine learning framework, we use sequence as well as diverse functional annotations to train models on multiple organisms using experimentally validated data. We present insights from tens of millions of features extracted and ranked from different modalities. We show high predictive performance per organism and in generalization across species. We provide a list of novel predictions including Danio rerio (zebrafish) and Arabidopsis thaliana (mouse-ear cress). We compare genomic models and observe that our protein model outperforms, whereas a unified model improves on both. While most membranous and disease related proteins are regulated by miRNAs, the G-protein coupled receptor (GPCR) family is an exception, being mostly unregulated by miRNAs. We further show that the evolutionary conservation among paralogs does not imply any coherence in miRNA regulation. We conclude that duplicated paralogous genes that often changed their function, also diverse in their tendency to be miRNA regulated. We conclude that protein function is informative across species in predicting post-transcriptional miRNA regulation in living cells.

show abstract

Section: Methodsmentioning

confidence: 99%

Inferring microRNA regulation: A proteome perspective

Ofer

Linial

2022

Front. Mol. Biosci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In fact, the patients admitted to two types of ICUs account for nearly 70% of the recorded total observations. Following recent work on measuring feature importance for predicting mortality [27] , [28] , we select the related variables and summarize their descriptions in Table 4 . All these variables are recognized as important predictors of mortality in patients.…”

Section: Experiments On the Icu Mortality Data In Us Hospitalsmentioning

confidence: 99%

Evaluating Risk-Adjusted Hospital Performance Using Large-Scale Data on Mortality Rates of Patients in Intensive Care Units: A Flexible Semi-Nonparametric Modeling Approach

Liang

Jiang

Zhang

2023

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

Background and objective: Significant variability in the quality of healthcare supplied by hospitals is drawing broad attention from the United States Centers for Medicare and Medicaid Services. The primary issue is to evaluate hospital performance based on patient outcomes. Generalized linear randomeffects models are a promising analytical tool for evaluating hospital performance. However, hospital compare data often violate the classical assumptions of normality on random effects and linearity representation on transformed conditional mean structures in these models. Methods: In this article, we proposed and tested the performance of a class of hospital compare models that embraces nonparametric mean structures with semi-nonparametric hospital random effects. Such models were further improved and integrated into a zero-inflated model. SAS programs to implement these newly proposed hospital compare models were thoroughly developed. The SAS programs are freely available via a GitHub (GitHub.com) repository. Results: We demonstrate the robustness of the proposed hospital compare models by conducting intensive empirical studies. Flexible seminonparametric random effects and functional fixed-effects mean structure were used to analyze patient mortality in a large-scale intensive care unit data set. After applying the proposed models to assess standardized modality rates and address patient-mix variability across hospitals, we detected those underperforming hospitals with higher mortality rates. Conclusions: Our research findings highlight how constructing advanced assessment tools for hospital performance could support better decision-making at the administrative and public levels. The proposed hospital compare models are comprehensive in their capacity to identify patterns of hospital random effects and to convey the variability in healthcare quality with powerful accuracy and interpretability.

show abstract

“…Test data augmentation can help increase the robustness of a trained model [ 73 , 74 , 75 ]. Test data augmentation may be utilized to enhance deep network prediction performance and open up new intriguing possibilities for medical image interpretation [ 76 , 77 , 78 ]. The most often utilized methods for augmenting data include rotating, mirroring, flipping, zooming, and cropping.…”

Section: Datasetsmentioning

confidence: 99%

LungNet22: A Fine-Tuned Model for Multiclass Classification and Prediction of Lung Disease Using X-ray Images

Shamrat

Azam

Karim

et al. 2022

JPM

View full text Add to dashboard Cite

In recent years, lung disease has increased manyfold, causing millions of casualties annually. To combat the crisis, an efficient, reliable, and affordable lung disease diagnosis technique has become indispensable. In this study, a multiclass classification of lung disease from frontal chest X-ray imaging using a fine-tuned CNN model is proposed. The classification is conducted on 10 disease classes of the lungs, namely COVID-19, Effusion, Tuberculosis, Pneumonia, Lung Opacity, Mass, Nodule, Pneumothorax, and Pulmonary Fibrosis, along with the Normal class. The dataset is a collective dataset gathered from multiple sources. After pre-processing and balancing the dataset with eight augmentation techniques, a total of 80,000 X-ray images were fed to the model for classification purposes. Initially, eight pre-trained CNN models, AlexNet, GoogLeNet, InceptionV3, MobileNetV2, VGG16, ResNet 50, DenseNet121, and EfficientNetB7, were employed on the dataset. Among these, the VGG16 achieved the highest accuracy at 92.95%. To further improve the classification accuracy, LungNet22 was constructed upon the primary structure of the VGG16 model. An ablation study was used in the work to determine the different hyper-parameters. Using the Adam Optimizer, the proposed model achieved a commendable accuracy of 98.89%. To verify the performance of the model, several performance matrices, including the ROC curve and the AUC values, were computed as well.

show abstract

ICU Survival Prediction Incorporating Test-Time Augmentation to Improve the Accuracy of Ensemble-Based Models

Cited by 25 publications

References 21 publications

Inferring microRNA regulation: A proteome perspective

Inferring microRNA regulation: A proteome perspective

Evaluating Risk-Adjusted Hospital Performance Using Large-Scale Data on Mortality Rates of Patients in Intensive Care Units: A Flexible Semi-Nonparametric Modeling Approach

LungNet22: A Fine-Tuned Model for Multiclass Classification and Prediction of Lung Disease Using X-ray Images

Contact Info

Product

Resources

About