DeepCov: Effective Prediction Model of COVID-19 Using CNN Algorithm

Abstract

COVID-19 outbreak prediction is a challenging and complicated problem in a vast dataset. Several communities have proposed various methods to predict the COVID-19-positive cases. However, conventional techniques remain drawbacks to predicting the actual trend cases. In this experiment, we adopt CNN to build our model by analyzing features from the vast COVID-19 dataset to predict long-term outbreaks to present early prevention. Our model can achieve adequate accuracy with a tiny loss based on the experiment results. In this study, we calculate the function which produces RMSE 0.00070 and MAPE 0.02440 to predict new cases and get RMSE 0.00468 and MAPE 0.06446 for predicting new deaths. Therefore, our proposed method can accurately predict the trend of positive cases in the COVID-19 outbreak.