Clinical predictions of COVID-19 patients using deep stacking neural networks

Chronic Kidney Disease (CKD) entails a progressive decline in renal function, often originating from comorbidities such as diabetes and hypertension. Early detection is crucial for managing progression and associated complications. Meanwhile, computed tomography (CT) serves as a vital tool for identifying kidney conditions. However, the accuracy and efficiency of radiological diagnoses are hampered by image quality, especially in early-stage disease. Despite extensive research on artificial intelligence (AI) based diagnostic models, high efficiency and optimal accuracy remains challenging. This study introduces a deep learning (DL) clinical diagnostic system aimed at enhancing the automatic identification and classification of CKD. Through an exploration of standard, advanced, and quantum DL algorithms in the CKD domain, it was selecting a hybrid quantum deep convolutional neural network (QDCNN) as our proposed model due to its high-quality performance. The model was optimized and refined through training with augmented and denoised datasets. This refined iteration yields impressive test performances in terms of accuracy: 99.98%, recall: 99.89%, precision: 99.84%, F1 score: 99.86%, specificity: 99.84%, Micro AUC: 99.99%, and testing time of 0.0641 seconds per image. Positioned to outperform existing methods, our proposed system demonstrates the potential to accurately identify kidney conditions at every stage, providing invaluable assistance to medical professionals and equipping them with an advanced level of accuracy, promptness, and reliability in diagnoses that is unparalleled in its excellence.

show abstract

Improving COVID-19 Mortality Predictions: A Stacking Ensemble Approach with Diverse Classifiers

Mohtasham,

Pourhoseingholi,

Nazari

et al. 2024

Preprint

View full text Add to dashboard Cite

Ensemble approaches are vital for developing effective machine learning methods by integrating multiple models to enhance performance and reduce bias and variance. This study utilized ensemble techniques to predict COVID-19 mortality using various classifiers. We first mapped the original dataset to a lower-dimensional space to improve training diversity. We then trained multiple base classifiers and ensemble methods, assessing their diversity through pairwise evaluations to create diverse combinations. A Stacking ensemble method was implemented with different meta-learners for improved predictive performance. All models were rigorously evaluated using standard discrimination and calibration metrics, along with statistical tests to identify significant performance differences. Various feature importance methods were applied to clarify the contributors to our model's predictions. The experimental results demonstrated the superiority of our stacking framework, specifically combining Random Forest and Extreme Gradient Boosting (XGBoost) with a Neural Network as the meta-learner on COVID-19 mortality prediction. This model achieved an accuracy of 0.914 (95% CI: 0.898, 0.928), precision of 0.818, F1-score of 0.801, Matthew’s correlation coefficient (MCC) of 0.746, and a ROC AUC of 0.955. These findings indicate that our framework is more effective than individual classifiers and existing ensemble methods, providing valuable insights for medical decision-making.

show abstract

Clinical predictions of COVID-19 patients using deep stacking neural networks

Cited by 4 publications

References 61 publications

Machine learning-based classification of valvular heart disease using cardiovascular risk factors

Machine learning-based classification of valvular heart disease using cardiovascular risk factors

Quantum Deep Learning for Automatic Chronic Kidney Disease Identification and Classification with CT images

Improving COVID-19 Mortality Predictions: A Stacking Ensemble Approach with Diverse Classifiers

Contact Info

Product

Resources

About