Advanced Forecasting of Covid-19 Infections Using Ensemble Deep Learning Methods

The global spread of the corona virus (COVID-19) is a pressing concern in the modern day. The effects of COVID-19 are noticeable throughout the population, and this includes the elderly who suffer from conditions such as asthma and other chronic illnesses. The medical industry has recently seen several accomplishments from ML and DL. Using ensemble deep learning models, this study primarily aims to forecast the figure of complete, fatal, and improved COVID-19 cases. Medical professionals can control the pandemic scenario with the use of accurate prediction models. The purpose of this learning was to determine the best prototypical for predicting the number of COVID-19 patients that survived, died, or recovered by using ensemble deep learning classifiers. Some of the approaches that are used are ResNet-50, Attention based LSTM (A-LSTM), and long short-term memory (LSTM). The Adaptive Bald Eagle Search Optimization (ABESO) program optimizes the neural networks' weights. The last step in predicting COVID-19 instances is to apply the softmax layer after ensembleing the network information via modality fusion. On two datasets, the WHO and JH, the collaborative method and current DL methods such as LSTM, ResNet-50, A-LSTM, and Bi-LSTM are evaluated using performance metrics like RMSE, MAE, and MAPE for confirmed, demise, and regaining cases.