COVID-19 County Level Severity Classification with Imbalanced Dataset: A NearMiss Under-sampling Approach

Oladunni, Timothy; Tossou, Sourou; Haile, Yayehyrad; Kidane, Adonias

doi:10.1101/2021.05.21.21257603

Cited by 4 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many recent studies have also attempted to predict COVID-19 patient clinical prognosis (either mortality, mechanical ventilation requirement, hospitalization or need for intubation) by feeding machine learning (ML) methods with clinical/demographic and/or radiomic features extracted from CXRs or HRCTs [3,[17][18][19][20][21][22][23][24][25]. In their recent study, Bae [17], Varghese [19], and Shiri [23] showed the potential usefulness of information extracted from radiographs.…”

Section: Sars-cov-2 Disease (Covid-19mentioning

confidence: 99%

Mortality Prediction of COVID-19 Patients Using Radiomic and Neural Network Features Extracted from a Wide Chest X-ray Sample Size: A Robust Approach for Different Medical Imbalanced Scenarios

et al. 2022

View full text Add to dashboard Cite

Aim: The aim of this study was to develop robust prognostic models for mortality prediction of COVID-19 patients, applicable to different sets of real scenarios, using radiomic and neural network features extracted from chest X-rays (CXRs) with a certified and commercially available software. Methods: 1816 patients from 5 different hospitals in the Province of Reggio Emilia were included in the study. Overall, 201 radiomic features and 16 neural network features were extracted from each COVID-19 patient’s radiography. The initial dataset was balanced to train the classifiers with the same number of dead and survived patients, randomly selected. The pipeline had three main parts: balancing procedure; three-step feature selection; and mortality prediction with radiomic features through three machine learning (ML) classification models: AdaBoost (ADA), Quadratic Discriminant Analysis (QDA) and Random Forest (RF). Five evaluation metrics were computed on the test samples. The performance for death prediction was validated on both a balanced dataset (Case 1) and an imbalanced dataset (Case 2). Results: accuracy (ACC), area under the ROC-curve (AUC) and sensitivity (SENS) for the best classifier were, respectively, 0.72 ± 0.01, 0.82 ± 0.02 and 0.84 ± 0.04 for Case 1 and 0.70 ± 0.04, 0.79 ± 0.03 and 0.76 ± 0.06 for Case 2. These results show that the prediction of COVID-19 mortality is robust in a different set of scenarios. Conclusions: Our large and varied dataset made it possible to train ML algorithms to predict COVID-19 mortality using radiomic and neural network features of CXRs.

show abstract

Section: Sars-cov-2 Disease (Covid-19mentioning

confidence: 99%

Mortality Prediction of COVID-19 Patients Using Radiomic and Neural Network Features Extracted from a Wide Chest X-ray Sample Size: A Robust Approach for Different Medical Imbalanced Scenarios

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The integration of NearMiss with PCA showed more accurate classifications. [34] evaluated two ensemble models and two non-ensemble models to predict county-level severity of COVID-19. NearMiss was used to undersample the majority class.…”

Section: Related Workmentioning

confidence: 99%

Evaluating undersampling techniques in the prediction of potential congenital syphilis cases using real data from Pernambuco, Brazil

Leite,

Rocha,

Teixeira

et al. 2024

Preprint

View full text Add to dashboard Cite

Syphilis can be transmitted congenitally and may cause serious consequences forthe child if not treated. The Programa Mãe Coruja Pernambucana (PMCP) is a brazilian public health program that helps pregnant women and saves data that can be used for prediction of potential congenital syphilis cases, through machine learning models. Only one work was found that predicts congenital syphilis through machine learning. This research uses a different methodology that evaluates undersampling in prediction. Random Undersampling, UnderSampling Based on Clustering (SBC) and NearMiss were used. The data was preprocessed and undersampling applied, generating different balanced datasets to train and test different machine learning models and different metrics for evaluation. Undersampling discarded data evaluation and analysis of distribution of the best attributes were applied to evaluate undersampling in best models. In models results, NearMiss trained models had high metrics, and very low in the discarded data. SBC models had smaller metrics, and in the discarded data went lower. Random Undersampling models had the lowest metrics, however in the discarded data showed similar results. The distribution of best attributes of NearMiss models were not similar to the original, contrary to Random Undersampling and SBC. NearMiss models had best results in the models, through this work evaluation showed that they cannot generalize the PMCP data and have not a representative distribution of the original data. Random Undersampling models had the lowest metrics but showed consistency through the evaluations, and thus are recommended for the congenital syphilis prediction.

show abstract

Learning from Imbalanced Data in Healthcare: State-of-the-Art and Research Challenges

Roy,

Roy

2024

Studies in Computational Intelligence

View full text Add to dashboard Cite

COVID-19 County Level Severity Classification with Imbalanced Dataset: A NearMiss Under-sampling Approach

Cited by 4 publications

References 20 publications

Mortality Prediction of COVID-19 Patients Using Radiomic and Neural Network Features Extracted from a Wide Chest X-ray Sample Size: A Robust Approach for Different Medical Imbalanced Scenarios

Mortality Prediction of COVID-19 Patients Using Radiomic and Neural Network Features Extracted from a Wide Chest X-ray Sample Size: A Robust Approach for Different Medical Imbalanced Scenarios

Evaluating undersampling techniques in the prediction of potential congenital syphilis cases using real data from Pernambuco, Brazil

Learning from Imbalanced Data in Healthcare: State-of-the-Art and Research Challenges

Contact Info

Product

Resources

About