Adaptive Deep Recurrent Neural Network-Based COVID-19 Healthcare Data Prediction for Early Risk Prediction

The Covid-19 pandemic has spread rapidly across the globe and is now one of the leading causes of death and illness worldwide. Existing approaches for controlling coronavirus disease are challenging because, the improper solutions, medications, and data are irregular to analyze. This paper proposes adaptive deep recurrent neural network-based Covid-19 healthcare data prediction, where the risk prediction algorithm is made to detect the Covid-19 disease when it is typically premature. The initially collected Covid-19 sample test dataset is trained in the preprocessing step to remove irrelevant data. The margins of features are estimated using threshold values to find the defect rate based on the Intrinsic Covid Defect Rate. The trained data are processed for feature selection using a threshold value to identify the best features using Relative Cluster-Intensive Feature Selection. The selected features are introduced to an Adaptive Deep vectorized Recursive Neural Network (ADVRNN) to predict the coronavirus affected rate. The results of the proposed ADVRNN experiment improve prediction accuracy, recall, f-measure, and precision rate to enhance the early detection prediction performance compared to the existing systems.