Time Series Analysis of Impact of Covid-19 Using Facebook Prophet Model and Review of the Machine-Learning Algorithm

The covid-19 pandemic has affected the entire world both financially and with loss of precious human life. This pandemic situation is not a new scenario as these kind of pandemic situations has affected us in past typically once in a century. Analysis the trend of covid-19 will helps us in understanding the requirement of the need to overcome this situation and it can used as a reference to tackle future pandemic of this kind. In this paper, the time series analysis of covid-19 is performed using facebook prophet model and the study of machine learning algorithm for covid-19 is performed. Facebook prophet model is a machine-learning algorithm introduced by facebook, which is very useful algorithm in predicting the future values for the time series data. In this article analysis of covid-19 data is carried out using Support Vector Machine (SVM), Random forest and Catboost algorithms and its results is studied followed by the time series analysis using facebook prophet model.

I. Introduction

Corona virus usually affects the animals and the birds and it is a member of the virus family named coronaviridae. This virus is classified into the following categories namely alpha, beta, gamma and delta. The occurrence of this virus was first spotted in the year 1930 in the chicken farm affecting the birds. It was in the year 1960, the virus affected the human affecting the respiratory track. The moderate effect of this virus caused common cold and the severity of this virus result in severe acute respiratory syndrome. In 2019, in the city of wuhan, china at the market the virus originated and the cases were only 44 in the first three months and started increasing rapidly from then onwards has spread all over the world. The only way to contain the spreading of this virus is early detection. The testing of this virus in human is carried out with polymerase chain reaction test also can be predicted using x ray and computer tomography images. The Convolutional Neural Networks plays a major role in dealing with these images in predicting the results of covid-19 and thereby assist the medical practitioner in taking the better decision [1]–[3]. Machine learning (ML) algorithms are also useful in predicting the covid -19 with the medical dataset. The next topic gives the application of ML in predicting covid-19 followed by the methodology. The next session gives a brief information about the ML algorithms such as SVM, Random Forest and catboost algorithm and the analysis of covi 19 data using these algorithms followed by the results and its analysis. And the last session gives the detailed use of time series analyses and the impact of the society in find the trend of covid-19 followed by the results obtained by facebook prophet model and the finally the conclusion and the future scope.