A Hybrid Deep Learning Model for Analyzing and Predicting Accurate Number of Patients with Health Conditions on Covid-19 Dataset

One of the recent real problems spreading like air pollution is Corona Virus Disease (Covid-19). The data generated by Covid-19 (C19) is crowdsourced. The epidemic of the C19 prevails as a challenge to world economic growth. In the future, the whole society meanwhile neither a curing medicine nor a preventing medicine needs to be discovered. The C19 spreads speedily time-to-time, increases sudden human death, and economy at risk. Including the world health organization, the health industries aim to reduce human extinction by detecting and preventing it at an earlier stage. Because of the increasing atrociousness of C19 cases, Hybrid Deep Learning (HDL) models are imperative in the current contexts. The DL model is the authoritative tool that fights against the C19 pandemic eruption by analyzing and predicting the severity of the cases early. The HDL model uses crowdsourced data to predict C19 cases worldwide in advance using Convolution Neural Network (CNN) and Long-Short-Term Memory (LSTM) for short?term, medium-term, and long-term dependencies. Before feeding the C19 crowdsourced data into the DL model, pre?process the data to improve prediction accuracy and reduce computational time and complexity. The CNN predicts the cumulative confirmation, deaths, cured cases, and the nested LSTM forecasts C19 cases in the HDL. This paper uses Python software to implement HDL to verify the performance metrics and evaluate the analyzed output to prove the model is reliable. The obtained results show that the HDL model yields high accuracy with the lowest error of 0.3%, which is efficient compared with the other existing models. Worldwide, country-wise, state-wise, and specific district-wise analysis reports of C19 data are given in detail, showing that the proposed HDL outperforms others.