ML Based Trend Analysis of Covid-19 in Pakistan

Abstract

COVID-19 is a virus having a family and comprised of different small to large viruses. The origin of this virus is Wuhan, the city of China. It was identified back in December 2019. COVID-19 spreads very rapidly in whole world. The spread of COVID-19 is different in every country due to geographical reasons; it is important to make a model that can explain each variant well. A system that predicts the diseases like corona virus before spreading out in the world. So, the precautionary steps can be taken on the results of system to stop or minimize the spreading of virus. A system which identifies the hot spot and cold spot areas of the corona virus. We uses the datasets of COVID-19 cases in Punjab province of Pakistan. We use different machine and deep learning models that are used for the prediction of time series data. In this study we use LSTM, ARIMA and Fbprofit for the prediction of covid cases. This paper also compares the results of these three models. We also identify the hot spot and cold spot districts of Punjab, so that the government can take precautionary steps in that region to control the spread of COVID-19. MAE and accuracy of these three models are used to figure out the best model among them. This research shows that the Prophet model is the best model among these to predict the COVID-19 cases in Punjab