Technologies to Confront the Coronavirus Crisis.
Algorithms, Paradigms, and Methodologies
In this article, I present a series of computer science, AI, data science machine learning, statistical analysis, and probability models in order to enhance the thinking and business landscape to face coronavirus.
Exhibit 1. Workspace setting at a hospital facility should align with data management as used in process management for agile decision-making.
The first problem that you face when trying to resolve a problem is that you need to figure out what resources you need and how much. Therefore, in principle, it is necessary to look at the current Coronavirus crisis not as an APEX-based problem, but rather as a queuing problem where servers have limited resources to maintain a robust service based on the queue size while leveraging a mean inter-arrival time based on real arrival rates, as in a Poisson process, and could otherwise make the queue[s] collapse, in which scenario both new cases and prevalence could grow exponentially without any possible control. Concurrent independent Poisson processes are possible in any coronavirus pandemic workspace. This is what the models created in the first few weeks predicted, in the lack of a cost reduction function [e.g., that created by social distancing in a disease gradient descent model] and other factors. This is also true because thinking about an APEX scenario works here as an assumption while thinking of a queuing model allows managers and leaders to take action on how to plan capacity (workspace) and resources, and the associated expenditure on resources and relevant operations. While this model is based on an average probability, it is capable to plan on resources for worst-case scenarios; the addition of a Navy boat as a large hospital is a good example, and imitating such a model, that is provisioning floating hospitals or eventually is a better choice than using hotels for this purpose, since perhaps one could think that setting up a large cruise boat as a temporary hospital is probably a much better choice in the long term. In the scenario of the City of New York, in particular, this is essentially appropriate and piers could serve as the workspace where ships could be attached as on-demand hospitals in any scenario for good. This probably provides better access to temporary hospitals through highways along the river piers. In a few words, reaching the APEX would allow medical scientists and researchers to adjust the queue-based management process rather than expecting for the APEX to occur and manage based on it. Utilizing this paradigm will also allow machine learning and data science expert to create simulations models for growth and survival analysis, and establish a consistent model that reflects the current status based on real live data on every stage.
Exhibit 2. The process management in the Coronavirus disease control should entail a clear alignment with the workspace for research data visualization, as such this exhibit simply depicts a Markov Chain of possible outcomes after all stages of a patient's treatment are completed.
Among the useful computer science technologies, I can include partitioning, algorithms as well as sorting and classification from AI, especially, reasoning under uncertainty and incomplete data, which are quite helpful to implement statistical trials that closely reflect the physical settings in a hospital. I believe that every nation should take specific actions, and think and plan in advance without panicking about a partitioning strategy that prevents the virus from spreading. For the time being, social distancing is just a basic model that has prevented the virus spread from going further.
The partitioning model should classify certain individuals as inexpugnable (really bolted) from the virus, meaning that their guaranteed to be virus-free in an extreme scenario, and this could include doctors, entities leaders, and managers, including both government and NGO leaders. This means that, for instance, under this model, British Prime Minister, Boris Johnson, and mission-critical specialized medical personnel should have never become ill with coronavirus. But most importantly, every citizen in a nation or other person regardless of their immigration status should be recognized as part of a partition both physically and logically. Likewise, partitioning techniques can also be used in the physical workspace of scientists and medical personnel. But most importantly it is relevant to create smart partitioning among patients by their statu quo from entry to remission. I may think about 5 categories, namely, Entry Level, Early Stage, Intermediate Stage, Critical Stage [SCU and ICU physical workspace], and Exit Stage, which includes either remission from the hospital or death itself. And this partitioning should also be used in a smart fashion for medical research and data scientists to work on issues such as signs and symptoms, disease evolution, prevalence, etc. In particular, physically and logically management the workspace is a consistent manner allows for better observation gathering, better data visualization, improved process management, and better heuristics from existent data. This can obviously entail the physical moving of patients throughout the physical workspace, i.e., the hospital installation, and it useful logical usage for data analysis and the further cohort of signs and symptoms, and further classification such as the impact on the race, gender, ethnicity, and other important demographics, with the usage of clustering techniques, and the analysis of sensitivity, specificity, and the finding of essential verifiable information about this disease. Furthermore, the usage of array-based workspace beyond partitioning could allow for the implementation of trials that allow researchers to use medications, such as hydroxychloroquine in early and intermediate coronavirus stages only or the vaccine the Bacille Calmette-Guerin vaccine (BCG), in optimal dosages in a customized fashion for every patient. This reminds me of my work close to the lead biostatistician and other researchers on treating crops with herbicides and insecticides from basic Latin squares arrays to more complex data classification settings for statistical testing or complex factorial design of experiments.
From machine learning models and methods, once data modeled based on physical and logical partitioning and clearly related to the management processes as wells the disease growth model[s], statistical analyses through both generative and discriminative [such as logistic regression] models or other supervised models and unsupervised models, such as those based on K-Means analysis, will be possible for management and government decision-making. Identifying patterns in any possible research category or dimension would allow researchers to deal with better self-questioning, analysis, and heuristics, and utilize preventive, curative [protective], and palliative healthcare, accordingly.
In summary, good management of the COVID-19 crisis should contemplate a queuing model, which is flexible and elastic, capable of providing workspace and resources on-demand rather than on the expectation of the APEX to occur at some point, because this is an uncertain event yet.
In essence, the closer the physical and logical model on the physical workspace (hospital) easily relates to and reassembles the models of the disease processes, the more appropriate strategic management will take place, allowing for the smartest usage and sharing of data.
Likewise, this overall paradigm will entail not only the improvement of future medications and vaccines but also the creation of educational methods that allow the average individual to learn and apply them accordingly. Thus, both the American and global economy will greatly benefit from the agile engineering of vaccines and medications for Coronavirus, and historic econometrics may suggest so.
Appendix
Reference and Useful Links
