Predictive modelling approach for cumulative fatality rate of COVID-19 in India
Main Article Content
Abstract
Abstract: Background: COVID-19 outbreak was first reported in Wuhan, China. In India, COVID-19 has 101139 laboratory-confirmed, 42,309 recovered and 3,163 deaths as on 19 May 2020. The study is conducted to select the best fit model for cumulative fatality rate due to COVID -19 in India from first death day to 19 May, 2020 and the predicted amount of cumulative fatality rate with respect to time is also presented. A comparison of different countries (where there are 50000+ cases registered by 19 May, 2020) has been prepared towards the average of cumulative fatality rate and average of cumulative recovery per death. Method: The goodness-of-fit tests: Kolmogorov-Smirnov test, Anderson-Darling test, Root Mean Square Error, and Coefficient of Determination are considered to select the best fit model for cumulative fatality rate. The probability distribution models which are frequently used in survival/death analysis are considered in the study. Results: On the behalf of the result of goodness-of-fit test, it is found that extreme value distribution model is selected as a best fit model on cumulative fatality rate due to COVID-19 in India. Conclusion: The average of cumulative fatality rate in United Kingdom is the highest and Russia has lowest. In terms of average of cumulative recovery per death, United Kingdom stands in the bottom and China is performing best. According to best fit model, the maximum value of cumulative fatality rate due to COVID-19 in India has been predicted as 4.2 and it is also predicting that the predicted value of cumulative fatality rate would be greater than 3.61, 3.65 and 3.7 after the 50, 80 and 110 days respectively.