Application of Mathematical Modeling in Prediction of COVID-19 Transmission Dynamics

Detection of COVID-19 epidemic outbreak using machine learning

Background

The coronavirus disease (COVID-19) pandemic has spread rapidly across the world, creating an urgent need for predictive models that can help healthcare providers prepare and respond to outbreaks more quickly and effectively, and ultimately improve patient care. Early detection and warning systems are crucial for preventing and controlling epidemic spread.

Objective

In this study, we aimed to propose a machine learning-based method to predict the transmission trend of COVID-19 and a new approach to detect the start time of new outbreaks by analyzing epidemiological data.

Methods

We developed a risk index to measure the change in the transmission trend. We applied machine learning (ML) techniques to predict COVID-19 transmission trends, categorized into three labels: decrease (L0), maintain (L1), and increase (L2). We used Support Vector Machine (SVM), Random Forest (RF), and XGBoost (XGB) as ML models. We employed grid search methods to determine the optimal hyperparameters for these three models. We proposed a new method to detect the start time of new outbreaks based on label 2, which was sustained for at least 14 days (i.e., the duration of maintenance). We compared the performance of different ML models to identify the most accurate approach for outbreak detection. We conducted sensitivity analysis for the duration of maintenance between 7 days and 28 days.

Results

ML methods demonstrated high accuracy (over 94%) in estimating the classification of the transmission trends. Our proposed method successfully predicted the start time of new outbreaks, enabling us to detect a total of seven estimated outbreaks, while there were five reported outbreaks between March 2020 and October 2022 in Korea. It means that our method could detect minor outbreaks. Among the ML models, the RF and XGB classifiers exhibited the highest accuracy in outbreak detection.

Conclusion

The study highlights the strength of our method in accurately predicting the timing of an outbreak using an interpretable and explainable approach. It could provide a standard for predicting the start time of new outbreaks and detecting future transmission trends. This method can contribute to the development of targeted prevention and control measures and enhance resource management during the pandemic.

Within-host delay differential model for SARS-CoV-2 kinetics with saturated antiviral responses

The present study discussed a model to describe the SARS-CoV-2 viral kinetics in the presence of saturated antiviral responses. A discrete-time delay was introduced due to the time required for uninfected epithelial cells to activate a suitable antiviral response by generating immune cytokines and chemokines. We examined the system's stability at each equilibrium point. A threshold value was obtained for which the system switched from stability to instability via a Hopf bifurcation. The length of the time delay has been computed, for which the system has preserved its stability. Numerical results show that the system was stable for the faster antiviral responses of epithelial cells to the virus concentration, i.e., quick antiviral responses stabilized patients' bodies by neutralizing the virus. However, if the antiviral response of epithelial cells to the virus increased, the system became unstable, and the virus occupied the whole body, which caused patients' deaths.

Mathematical modeling for Delta and Omicron variant of SARS-CoV-2 transmission dynamics in Greece

A compartmental, epidemiological, mathematical model was developed in order to analyze the transmission dynamics of Delta and Omicron variant, of SARS-CoV-2, in Greece. The model was parameterized twice during the 4th and 5th wave of the pandemic. The 4th wave refers to the period during which the Delta variant was dominant (approximately July to December of 2021) and the 5th wave to the period during which the Omicron variant was dominant (approximately January to May of 2022), in accordance with the official data from the National Public Health Organization (NPHO). Fitting methods were applied to evaluate important parameters in connection with the transmission of the variants, as well as the social behavior of population during these periods of interest. Mathematical models revealed higher numbers of contagiousness and cases of asymptomatic disease during the Omicron variant period, but a decreased rate of hospitalization compared to the Delta period. Also, parameters related to the behavior of the population in Greece were also assessed. More specifically, the use of protective masks and the abidance of social distancing measures. Simulations revealed that over 5,000 deaths could have been avoided, if mask usage and social distancing were 20% more efficient, during the short period of the Delta and Omicron outbreak. Furthermore, the spread of the variants was assessed using viral load data. The data were recorded from PCR tests at 417 Army Equity Fund Hospital (NIMTS), in Athens and the Ct values from 746 patients with COVID-19 were processed, to explain transmission phenomena and disease severity in patients. The period when the Delta variant prevailed in the country, the average Ct value was calculated as 25.19 (range: 12.32-39.29), whereas during the period when the Omicron variant prevailed, the average Ct value was calculated as 28 (range: 14.41-39.36). In conclusion, our experimental study showed that the higher viral load, which is related to the Delta variant, may interpret the severity of the disease. However, no correlation was confirmed regarding contagiousness phenomena. The results of the model, Ct analysis and official data from NPHO are consistent.

Enhancing biofeedback-driven self-guided virtual reality exposure therapy through arousal detection from multimodal data using machine learning

Virtual reality exposure therapy (VRET) is a novel intervention technique that allows individuals to experience anxiety-evoking stimuli in a safe environment, recognise specific triggers and gradually increase their exposure to perceived threats. Public-speaking anxiety (PSA) is a prevalent form of social anxiety, characterised by stressful arousal and anxiety generated when presenting to an audience. In self-guided VRET, participants can gradually increase their tolerance to exposure and reduce anxiety-induced arousal and PSA over time. However, creating such a VR environment and determining physiological indices of anxiety-induced arousal or distress is an open challenge. Environment modelling, character creation and animation, psychological state determination and the use of machine learning (ML) models for anxiety or stress detection are equally important, and multi-disciplinary expertise is required. In this work, we have explored a series of ML models with publicly available data sets (using electroencephalogram and heart rate variability) to predict arousal states. If we can detect anxiety-induced arousal, we can trigger calming activities to allow individuals to cope with and overcome distress. Here, we discuss the means of effective selection of ML models and parameters in arousal detection. We propose a pipeline to overcome the model selection problem with different parameter settings in the context of virtual reality exposure therapy. This pipeline can be extended to other domains of interest where arousal detection is crucial. Finally, we have implemented a biofeedback framework for VRET where we successfully provided feedback as a form of heart rate and brain laterality index from our acquired multimodal data for psychological intervention to overcome anxiety.

EEI-IoT: Edge-Enabled Intelligent IoT Framework for Early Detection of COVID-19 Threats

Coronavirus disease 2019 (COVID-19) has caused severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) across the globe, impacting effective diagnosis and treatment for any chronic illnesses and long-term health implications. In this worldwide crisis, the pandemic shows its daily extension (i.e., active cases) and genome variants (i.e., Alpha) within the virus class and diversifies the association with treatment outcomes and drug resistance. As a consequence, healthcare-related data including instances of sore throat, fever, fatigue, cough, and shortness of breath are given due consideration to assess the conditional state of patients. To gain unique insights, wearable sensors can be implanted in a patient's body that periodically generates an analysis report of the vital organs to a medical center. However, it is still challenging to analyze risks and predict their related countermeasures. Therefore, this paper presents an intelligent Edge-IoT framework (IE-IoT) to detect potential threats (i.e., behavioral and environmental) in the early stage of the disease. The prime objective of this framework is to apply a new pre-trained deep learning model enabled by self-supervised transfer learning to build an ensemble-based hybrid learning model and to offer an effective analysis of prediction accuracy. To construct proper clinical symptoms, treatment, and diagnosis, an effective analysis such as STL observes the impact of the learning models such as ANN, CNN, and RNN. The experimental analysis proves that the ANN model considers the most effective features and attains a better accuracy (~98.3%) than other learning models. Also, the proposed IE-IoT can utilize the communication technologies of IoT such as BLE, Zigbee, and 6LoWPAN to examine the factor of power consumption. Above all, the real-time analysis reveals that the proposed IE-IoT with 6LoWPAN consumes less power and response time than the other state-of-the-art approaches to infer the suspected victims at an early stage of development of the disease.

Four-way classification of Alzheimer's disease using deep Siamese convolutional neural network with triplet-loss function

Alzheimer's disease (AD) is a neurodegenerative disease that causes irreversible damage to several brain regions, including the hippocampus causing impairment in cognition, function, and behaviour. Early diagnosis of the disease will reduce the suffering of the patients and their family members. Towards this aim, in this paper, we propose a Siamese Convolutional Neural Network (SCNN) architecture that employs the triplet-loss function for the representation of input MRI images as k-dimensional embeddings. We used both pre-trained and non-pretrained CNNs to transform images into the embedding space. These embeddings are subsequently used for the 4-way classification of Alzheimer's disease. The model efficacy was tested using the ADNI and OASIS datasets which produced an accuracy of 91.83% and 93.85%, respectively. Furthermore, obtained results are compared with similar methods proposed in the literature.

A Mathematical Model of Vaccinations Using New Fractional Order Derivative

Purpose: This paper studies a simple SVIR (susceptible, vaccinated, infected, recovered) type of model to investigate the coronavirus’s dynamics in Saudi Arabia with the recent cases of the coronavirus. Our purpose is to investigate coronavirus cases in Saudi Arabia and to predict the early eliminations as well as future case predictions. The impact of vaccinations on COVID-19 is also analyzed. Methods: We consider the recently introduced fractional derivative known as the generalized Hattaf fractional derivative to extend our COVID-19 model. To obtain the fitted and estimated values of the parameters, we consider the nonlinear least square fitting method. We present the numerical scheme using the newly introduced fractional operator for the graphical solution of the generalized fractional differential equation in the sense of the Hattaf fractional derivative. Mathematical as well as numerical aspects of the model are investigated. Results: The local stability of the model at disease-free equilibrium is shown. Further, we consider real cases from Saudi Arabia since 1 May−4 August 2022, to parameterize the model and obtain the basic reproduction number R0v≈2.92. Further, we find the equilibrium point of the endemic state and observe the possibility of the backward bifurcation for the model and present their results. We present the global stability of the model at the endemic case, which we found to be globally asymptotically stable when R0v>1. Conclusion: The simulation results using the recently introduced scheme are obtained and discussed in detail. We present graphical results with different fractional orders and found that when the order is decreased, the number of cases decreases. The sensitive parameters indicate that future infected cases decrease faster if face masks, social distancing, vaccination, etc., are effective.