Predictive Analysis of COVID-19 Data Using Two-Step Quantile Regression Method

ABSTRACT

In the year 2019, research community began with new challenge called novel coronavirus disease (COVID-19) and has opened up new challenges for the research community. From the report of the World Health Organization (WHO), the new virus COVID-2019 (World Health Organization (2020) Coronavirus disease 2019 (COVID-19) situation report, p 67) causes dangerous illness to the concerned person, and it spread to other peoples with huge rate through contact. Such kind of pandemic analysis needs efficient methods to predict data and also helped further to analyze such epidemic risks. These kinds of analyses are used to handle and control the epidemic kind of diseases. Regression analysis is kind of ML methods and is worked well to analyze such kind of epidemic data. The work in this paper about analysis of COVID-19 data especially focused in the state of Andhra Pradesh. First, the data are collected from the website (i.e., https//prsindia.org/ ). Next, we applied various regression techniques like linear, multi-linear and quantile regression for COVID-19 data for the prediction of cases. Further extended work to derive penalized quantile results using lasso. The results shows that the two-step quantile regression (TSQR) has been shown to be a better predictive method for predicting confirmed cases compared to linear and multi-linear regressions in terms of MSE and R-Score parameters. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.