COVID-19 Outbreak Prediction with Machine Learning

Abstract

Several outbreak prediction models for COVID-19 are being used by officials around the world to make informed decisions and enforce relevant control measures. Among the standard models for COVID-19 global pandemic prediction, simple epidemiological and statistical models have received more attention by authorities, and these models are popular in the media. Due to a high level of uncertainty and lack of essential data, standard models have shown low accuracy for long-term prediction. Although the literature includes several attempts to address this issue, the essential generalization and robustness abilities of existing models need to be improved. This paper presents a comparative analysis of machine learning and soft computing models to predict the COVID-19 outbreak as an alternative to susceptible–infected–recovered (SIR) and susceptible-exposed-infectious-removed (SEIR) models. Among a wide range of machine learning models investigated, two models showed promising results (i.e., multi-layered perceptron, MLP; and adaptive network-based fuzzy inference system, ANFIS). Based on the results reported here, and due to the highly complex nature of the COVID-19 outbreak and variation in its behavior across nations, this study suggests machine learning as an effective tool to model the outbreak. This paper provides an initial benchmarking to demonstrate the potential of machine learning for future research. This paper further suggests that a genuine novelty in outbreak prediction can be realized by integrating machine learning and SEIR models.

A Novel Model for Simulating COVID-19 Dynamics Through Layered Infection States that Integrate Concepts from Epidemiology, Biophysics and Medicine: SEI3R2S-Nrec

“…Outbreaks were more regional, and directives diverse. Further, early "lockdown" directives, while aggressive, were suboptimal, as organizations such as WHO had CDC hadn't yet acknowledged the scope of airborne aerosol transmission and essential nature of mask use (27)(28)(29)(30)(31)63).…”

“…Both conceptualizations have merit, but each views an "exposure event" differently. It was biophysical analysis dating back to March that led to the recognition that a key transmission mode of viral particles is in the form of aerosols (1,2,(28)(29)(30)(31)(32)(33)(34)54,55), and associated with this, the overriding importance of mask use and of limiting indoor group gatherings (27)(28)(29)(30)(31)(32)(33)(34), and less so outdoor gatherings such as protests with high mass use (56).…”

“…Studies have also considered roles of modern transportation on world-wide infection spread via host translocation operating through emergent "hub-spoke" "small world" complexity models (27) where "hubs" can be brain regions/cities/airports/web-apps and "spokes" involve local spread that evolves in space and time (25,28). Some "metapopulation" models have even mapped geospatial data to the types of facilities where people gather and the contexts associated with such encounters (25), or estimated risks of local gatherings based in part on emerging serology data (29, https://covid19risk.biosci.gatech.edu/), or used machine learning/neurofuzzy tools (30). Such models and tools, while intriguing and of value, may be limited as they infer rather than directly document the form of actual encounters between people, and tend not to be able to address important contextual information such as mask use (but could help drive SEIRS models).…”

“…Actual drives are clearly the fruits of manifold agent-based CAS processes, both within society and within our bodies. Valuable studies of social behavior as related to COVID-19 range from mobility (23)(24)(25)(26)(27)(28)(29)(30) to impacts of mask use (2,(32)(33)(34)(35)(36)(37)(38). All have merit, and such research was are not intended to help drive an epidemiologicallyand medically-oriented model.…”

Identification of high-risk COVID-19 patients using machine learning

“…Undoubtedly, the ability to rapidly identify high-risk patients and/or correctly assign health care priorities is critical, in the first case so as to improve hospital capacity planning and in the second case for providing timely treatment for patients [16]. In this regard, artificial intelligence methods have been recognized as a powerful and promising technology that can help not only in the identification of the fatality risk of a given patient seeking medical attention [17,18], but also for the diagnosis process [19][20][21][22], prediction of disease spreading dynamics [23][24][25][26][27], and tracking of infected patients as well as likely future patients [28].…”