Estimate the incubation period of coronavirus 2019 (COVID-19)

Computers in Biology and Medicine, Год журнала: 2023, Номер 158, С. 106794 - 106794

Опубликована: Март 30, 2023

Язык: Английский

---

LSTM algorithm optimization for COVID-19 prediction model

Heliyon, Год журнала: 2024, Номер 10(4), С. e26158 - e26158

Опубликована: Фев. 1, 2024

The development of predictive models for infectious diseases, specifically COVID-19, is an important step in early control efforts to reduce the mortality rate. However, traditional time series prediction used analyze disease spread trends often encounter challenges related accuracy, necessitating need develop with enhanced accuracy. Therefore, this research aimed a model based on Long Short-Term Memory (LSTM) networks better predict number confirmed COVID-19 cases. proposed optimized LSTM (popLSTM) was compared Basic and improved MinMaxScaler developed earlier using dataset taken from previous research. collected four countries high daily increase cases, including Hong Kong, South Korea, Italy, Indonesia. results showed significantly accuracy methods. contributions popLSTM included 1) Incorporating output gate effectively filter more detailed information model, 2) Reducing error value by considering hidden state improve experiment exhibited significant 4%

Язык: Английский

---

A novel hybrid supervised and unsupervised hierarchical ensemble for COVID-19 cases and mortality prediction

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Апрель 29, 2024

Abstract Though COVID-19 is no longer a pandemic but rather an endemic, the epidemiological situation related to SARS-CoV-2 virus developing at alarming rate, impacting every corner of world. The rapid escalation coronavirus has led scientific community engagement, continually seeking solutions ensure comfort and safety society. Understanding joint impact medical non-medical interventions on spread essential for making public health decisions that control pandemic. This paper introduces two novel hybrid machine-learning ensembles combine supervised unsupervised learning data classification regression. study utilizes publicly available outbreak potential predictive features in USA dataset, which provides information disease US, including from each 3142 US counties beginning epidemic (January 2020) until June 2021. developed hierarchical classifiers outperform single algorithms. best-achieved performance metrics task were Accuracy = 0.912, ROC-AUC 0.916, F1-score 0.916. proposed ensemble combining both allows us increase accuracy regression by 11% terms MSE, 29% area under ROC, 43% MPP metric. Thus, using approach, it possible predict number cases deaths based demographic, geographic, climatic, traffic, health, social-distancing-policy adherence, political characteristics with sufficiently high accuracy. reveals pressure most important feature analysis. Five other significant identified have influence spread. combined ensembling approach introduced this can help policymakers design prevention measures avoid or minimize threats future.

Язык: Английский

---

Deep Learning for Covid-19 Identification: A Comparative Analysis

International Research Journal on Advanced Science Hub, Год журнала: 2022, Номер 4(11), С. 272 - 280

Опубликована: Ноя. 28, 2022

Covid 19 was an epidemic in 2022. Detection of X-Ray samples is crucial for diagnosis and treatment. This also challenging the identification covid by radiologists. study proposes Transfer Learning detecting Covid-19 from images. The proposed detects normal x-ray samples. In addition to this model, different architectures including trained Desnet121, Efficient B4, Resnet 34, mobilenetv2 were evaluated dataset. Our suggested model has compared existing covid-19 detection algorithm terms accuracy. Experimental patients with accuracy 98 percent. work analyse covid19 automation helps deep learning algorithms which results high Covid19 using can assist radiologists doctors make test more accessible.

Язык: Английский

---

Predicting hospital readmission risk in patients with COVID-19: A machine learning approach

Informatics in Medicine Unlocked, Год журнала: 2022, Номер 30, С. 100908 - 100908

Опубликована: Янв. 1, 2022

The Coronavirus 2019 (COVID-19) epidemic stunned the health systems with severe scarcities in hospital resources. In this critical situation, decreasing COVID-19 readmissions could potentially sustain capacity. This study aimed to select most affecting features of readmission and compare capability Machine Learning (ML) algorithms predict based on selected features. data 5791 hospitalized patients were retrospectively recruited from a registry system. LASSO feature selection algorithm was used important related readmission. HistGradientBoosting classifier (HGB), Bagging classifier, Multi-Layered Perceptron (MLP), Support Vector ((SVM) kernel = linear), SVM (kernel RBF), Extreme Gradient Boosting (XGBoost) classifiers for prediction. We evaluated performance ML 10-fold cross-validation method using six evaluation metrics. Out 42 features, 14 identified as relevant predictors. XGBoost outperformed other models an average accuracy 91.7%, specificity 91.3%, sensitivity 91.6%, F-measure 91.8%, AUC 0.91%. experimental results prove that can satisfactorily Besides considering risk factors prioritized work, categorizing cases high reinfection make patient triaging procedure resource utilization more effective.

Язык: Английский

---

Short-term forecasting of COVID-19 using support vector regression: An application using Zimbabwean data

American Journal of Infection Control, Год журнала: 2023, Номер 51(10), С. 1095 - 1107

Опубликована: Март 30, 2023

Язык: Английский

---

An Optimized Model Based on Deep Learning and Gated Recurrent Unit for COVID-19 Death Prediction

Biomimetics, Год журнала: 2023, Номер 8(7), С. 552 - 552

Опубликована: Ноя. 17, 2023

The COVID-19 epidemic poses a worldwide threat that transcends provincial, philosophical, spiritual, radical, social, and educational borders. By using connected network, healthcare system with the Internet of Things (IoT) functionality can effectively monitor cases. IoT helps patient recognize symptoms receive better therapy more quickly. A critical component in measuring, evaluating, diagnosing risk infection is artificial intelligence (AI). It be used to anticipate cases forecast alternate incidences number, retrieved instances, injuries. In context COVID-19, technologies are employed specific monitoring processes reduce exposure others. This work uses an Indian dataset create enhanced convolutional neural network gated recurrent unit (CNN-GRU) model for death prediction via IoT. data were also subjected normalization imputation. 4692 eight characteristics utilized this research. performance CNN-GRU was assessed five evaluation metrics, including median absolute error (MedAE), mean (MAE), root squared (RMSE), square (MSE), coefficient determination (R2). ANOVA Wilcoxon signed-rank tests determine statistical significance presented model. experimental findings showed outperformed other models regarding prediction.

Язык: Английский

---

Improved healthcare disaster decision-making utilizing information extraction from complementary social media data during the COVID-19 pandemic

Decision Support Systems, Год журнала: 2023, Номер 172, С. 113983 - 113983

Опубликована: Апрель 24, 2023

Язык: Английский

---

A probabilistic spatio-temporal neural network to forecast COVID-19 counts

International Journal of Data Science and Analytics, Год журнала: 2024, Номер unknown

Опубликована: Март 21, 2024

Abstract Geo-referenced and temporal data are becoming more ubiquitous in a wide range of fields such as medicine economics. Particularly the realm medical research, spatio-temporal play pivotal role tracking understanding spread dynamics diseases, enabling researchers to predict outbreaks, identify hot spots, formulate effective intervention strategies. To forecast these types we propose Probabilistic Spatio-Temporal Neural Network that (1) estimates, with computational efficiency, models spatial components; (2) combines flexibility Network—which is free from distributional assumptions—with uncertainty quantification probabilistic models. Our architecture compared established INLA method, well other baseline models, on COVID-19 Italian regions. empirical analysis demonstrates superior predictive effectiveness our method across multiple ranges offers insights for shaping targeted health interventions

Язык: Английский

---

Interpretation of COVID-19 Epidemiological Trends in Mexico Through Wastewater Surveillance Using Simple Machine Learning Algorithms for Rapid Decision-Making

Viruses, Год журнала: 2025, Номер 17(1), С. 109 - 109

Опубликована: Янв. 15, 2025

Detection and quantification of disease-related biomarkers in wastewater samples, denominated Wastewater-based Surveillance (WBS), has proven a valuable strategy for studying the prevalence infectious diseases within populations time- resource-efficient manner, as samples are representative all cases catchment area, whether they clinically reported or not. However, analysis interpretation WBS datasets decision-making during public health emergencies, such COVID-19 pandemic, remains an area opportunity. In this article, database obtained from sampling at treatment plants (WWTPs) university campuses Monterrey Mexico City between 2021 2022 was used to train simple clustering- regression-based risk assessment models allow informed prevention control measures high-affluence facilities, even if working with low-dimensionality limited number observations. When dividing weekly data points based on seven-day average daily new were above certain threshold, resulting clustering model could differentiate weeks surges clinical reports periods them 87.9% accuracy rate. Moreover, provided satisfactory forecasts one week (80.4% accuracy) two (81.8%) into future. prediction (R2 = 0.80, MAPE 72.6%), likely because insufficient dimensionality database. Overall, while simple, WBS-supported can provide relevant insights decision-makers epidemiological outbreaks, regression algorithms using still be improved.

Язык: Английский

---