Machine Learning and Deep Learning Based Time Series Prediction and Forecasting of Ten Nations' COVID-19 Pandemic.

In the paper, the authors investigated and predicted the future environmental circumstances of a COVID-19 to minimize its effects using artificial intelligence techniques. The experimental investigation of COVID-19 instances has been performed in ten countries, including India, the United States, Russia, Argentina, Brazil, Colombia, Italy, Turkey, Germany, and France using machine learning, deep learning, and time series models. The confirmed, deceased, and recovered datasets from January 22, 2020, to May 29, 2021, of Novel COVID-19 cases were considered from the Kaggle COVID dataset repository. The country-wise Exploratory Data Analysis visually represents the active, recovered, closed, and death cases from March 2020 to May 2021. The data are pre-processed and scaled using a MinMax scaler to extract and normalize the features to obtain an accurate prediction rate. The proposed methodology employs Random Forest Regressor, Decision Tree Regressor, K Nearest Regressor, Lasso Regression, Linear Regression, Bayesian Regression, Theilsen Regression, Kernel Ridge Regressor, RANSAC Regressor, XG Boost, Elastic Net Regressor, Facebook Prophet Model, Holt Model, Stacked Long Short-Term Memory, and Stacked Gated Recurrent Units to predict active COVID-19 confirmed, death, and recovered cases. Out of different machine learning, deep learning, and time series models, Random Forest Regressor, Facebook Prophet, and Stacked LSTM outperformed to predict the best results for COVID-19 instances with the lowest root-mean-square and highest R 2 score values.

Decompensation Prediction for Hospitalized COVID-19 Patients

Since the start of the COVID-19 pandemic, hospitals have been overwhelmed with the high number of ill and critically ill patients. The surge in ICU demand led to ICU wards running at full capacity, with no signs of demand falling. As a result, resource management of ICU beds and ventilators has been a bottleneck in providing adequate healthcare to those in need. Short-term ICU demand forecasts have become a critical tool for hospital administrators. Therefore, using the existing COVID-19 patient data, we build models to predict if a patient's health will deteriorate below safe thresholds to deem admission into ICU in the next 24 to 96 hours. We identify the most important clinical features responsible for the prediction and narrow down the health indicators to focus on, thereby assisting the hospital staff in increasing responsiveness. These models can help the hospital staff better forecast ICU demand in near real-time and triage patients for ICU admissions as per the risk of deterioration. Using a retrospective study with a dataset of 1411 COVID-19 patients from an actual hospital in the USA, we run experiments and find XGBoost performs the best among the models tested when tuning parameters for sensitivity (recall). The most important feature for the four prediction tasks is the maximum respiratory rate, but subsequent features in order of importance vary between models predicting ICU transfer in the next 24 to 48 hours and those predicting ICU transfer in the next 72 to 96 hours. © 2022 IEEE.

Ensemble method for multiclassification of COVID-19 virus using spatial and frequency domain features over X-ray images

In early 2020, WHO declared COVID-19, a pandemic disease, which severely infected human inhabitant and health. Researchers, doctors, etc., are finding ways to combat the disease. RT-PCR testing is the initial type of testing that was used to detect whether a patient is COVID (+) or COVID (−).This test kit is costly and the result takes around 6hours. So testing a heavy chunk of the population with RT-PCR is a difficult task. To counter this, X-rays/CT scan-based testing can be used to detect COVID (+) cases to control its spread. X-rays are preferable to CT as they are cheaper and even produce low radiations. The second issue that was noticed during this pandemic period was the availability of doctors. To resolve this issue, a robust automated system for early prediction is essential. Automated systems using machine learning (ML), deep learning (DL) approaches are giving promising results in the detection of COVID (+) cases. In this chapter, we propose a framework for automatic recognition of COVID (+), normal, and pneumonia cases (i.e., multiclassification) over X-ray images. In the proposed method, a dataset of COVID (+), normal, and pneumonia images is used. Initially, the dataset is preprocessed, followed by feature extraction using gray level cooccurrence matrix (GLCM), gray level difference method (GLDM), wavelet transform (WT), and fast Fourier transform (FFT) methods. Features extracted are concatenated to construct a feature pool and these features are used for multiclassification using ML algorithms: support vector machines (SVM) and XG Boost. XG Boost performs better than SVM. © 2022 Elsevier Inc. All rights reserved.