Analysis of origin, risk factors influencing COVID-19 cases in India and its prediction using ensemble learning

Abstract

A genetically heterogeneous infectious virus recognized as COVID-19 (COronaVIrus Disease) has a significant influence on worldwide health. COVID-19 is an infection transmitted by a viral disease known as Severe Acute Respiratory Syndrome CoronaVirus-2. (SARS-CoV-2). In the current period, the instances of COVID-19 outbreaks are spreading significantly over the globe. Swallowed (through inhalation) or contact with contaminated droplets transmits the illness. Symptoms appear from two to fourteen days after exposure . Many individuals have a symptomless infection. The emphasis of therapy is on fundamental human needs; further research is needed on the risk analysis of the disease and use of antiviral medications. This research focuses on the case instances through the chronological dates in various parts of Indian states. The primary analysis of case instance is dependent on three stages; they are confirming stage, fatality stage, and recovered stage. The relevant datasets are obtained through Kaggle (from January 2020 to May 2021). The dataset has multiple categories that further intensify to perform high-end classifications. Initially, data preprocessing is done that include cleaning and feature selection on the prepared dataset, and then the prediction process is carried out using an Ensemble training approach. Approaches of Ensemble learning utilize many learning algorithms to achieve improved prediction efficiency. The stacking mechanism is used for aggregating accuracy. From the experimental observations, the suggested learning strategy is found to attain the maximum degree of precision for each stage (confirmed: 84.37%, fatality: 82.13%, and recovered: 88.67%) compared to the other approaches that are chosen to perform alone. Assessing the stability of the learning model is done by bootstrapping validation.