How COVID19 pandemic has been tackled by Data Science & Machine Learning

                             

Photo by Ashkan Forouzani on Unsplash

Machine learning and data science have been playing a vital role in the fight against the recent deadly pandemic known as COVID19. Almost all countries have been using some way or the other this data science, machine learning, and other means of statistical analysis to see the growth rate of the curve of the infected patients and to realize the death rate caused due to this pandemic.

                                            

However, there are a few challenges in this field and it is high time that we address those issues. 

The prime struggle in the data science applicability towards covid19 is collecting the data. This seems to be one of the most challenging problems that all data scientists are facing right now. One aspect of it of getting numerical data to predict the curve of death and affected people. The other application of machine learning is to apply the deep learning and advanced machine learning to classify X-Ray images of COVID19 patients and normal pneumonia patients. The CT scan images of COVID19 patients are also available for machine learning and deep learning applications. To make these applications successful one needs high-end powerful processors that may not be available in this hard time to all countries' medical scientists.

Besides, data science and machine learning can help to do the following performances as well

1. It can help to discover drug that is the need of this hour;  drug discovery technology can be uplifted with the use of machine learning and deep learning.

2. It can help decide the red zone, orange zone, and green zone based on the density of the infected patients.

3. It can do the feature extractions from the images of covid19 patients to understand the behavior of the virus.

4. It can also do justification towards the source from where the virus has been originated; that will help to estimate the next big outbreak.

Most significantly the data science and machine learning can help to segregate the patients based on the severity of the infection so that doctors can decide whom to keep quarantine, and who to provide proper medications or treatments.

It can predict to some extent with some threshold, on the possible outcome of the treatment. The geographic segregation based on the locality in terms of country wise, state wise, district wise, and then finally as continent wise are also possible using machine learning and data science. For example, the COVID19 pandemic has been more severe in European countries and in the USA than in Asia, Australia. etc. Using the pattern from this data the experts can normalize it so that a  generic statement can be built related to the growth of infection rate, death rate, and a decision that can be made when the graph of COVID patient shall be flattening (possibility)?

As I mentioned the total death can also be predicted; as experts recently predicted for the UK that the death toll will be even higher than 66,000. According to the washingtonpost, on 25th April's post, President Trump said Friday(24th April) that he expects the total number of U.S. deaths, once the pandemic has abated, to be “hopefully far below” early minimum estimates of 100,000. On the same Washington Post, it has been written that: "The U.S. death toll from the coronavirus surpassed 51,000 on Friday(24th April), more than a quarter of all the virus-related deaths reported worldwide and quickly approaching the number of Americans — 58,200 — who died during the Vietnam War". 

In this pandemic time, deep learning can also be a useful tool as it may able to analyze the facial expression by the facial screening of infected patients(This may be a possibility).

The other significant point is to ignore the outlier from the data sets such as image and numeric data related to covid19 patients and infection rates. However, it needs a piece of expert knowledge.

Lastly, the data science expert must build their model keeping in mind that it can not be completely perfect. They have to consider their models with some reservations. The catch is that these models are just helping hands towards the successful treatment of  COVID19 infection.

We hope that we will very soon achieve success by eliminating coronavirus cases from the earth using multi-dimensional treatments including data science and machine learning. Let us pray to god together.

~AISavvy


You can read this story also from  #Medium.

Read these following blogposts also,
1. Indian workforce shows the best digital skills in the world: Gartner Survey
2. How Covid19 has already impacted the Data Science JOB Market