Deep adaptive CHIONet: designing novel herd immunity prediction of COVID-19 pandemic using hybrid RNN with LSTM

The rapid spread of COVID-19 threatened the entire world because of its adverse effects and high mortality rate. The most effective method of disease prevention is immunisation and vaccination. The development of herd immunity against any deadly virus will stop the pandemic. The main goal of this research is to develop an enhanced herd immunity prediction model for the COVID-19 pandemic. To develop a prediction model, a hybrid RNN and LSTM are combined with the proposed ACHIO. Feature extraction and feature selection methods are used to select the most important features that enhance the model’s performance. Once the features are extracted using statistical methods, the optimal feature selection is performed by ACHIO. The selected features are then fed into the RNN and LSTM, and the epoch and neuron count in the RNN and LSTM is optimised using ACHIO to improve model performance. The proposed model achieved 90.42% accuracy, 80% precision, 90.86% specificity, 89.53% sensitivity, 86.03% F1-Score, 17.20% FDR, 90.86% NPV, 10.47% FNR, and 9.14% FPR. Various deep learning models, including DNN, RNN, CNN, RBM, LSTM, and RNN + LSTM, are compared to evaluate the performance of the proposed model. The results indicate that the proposed model performs better than the existing standard.