Disease forecasting and severity prediction model for COVID-19 using correlated feature extraction and feed-forward artificial neural networks

Jayaraj, T.; Samath, J. A.

The digitization of the medical sector has led to an explosion of heterogeneous medical records. The contribution of big data in the medical field is used to effectively address certain unsolved issues. Effectively integrating and analyzing this big data can reveal many useful hidden medical information. The COVID-19 virus, which first appeared in China at the end of 2019, is suffocating the rest of the world. Traditional methods of preventing this unforeseen pandemic are the lengthy process that can take several years. Now, the second wave of COVID-19 is wreaking havoc throughout the world. The survival rate can be significantly increased by predicting the risk of COVID-19 infection based on the patient's early symptoms and health status. Various disease forecasting approaches based on machine learning, and deep learning has been developed to forecast the severity rate of COVID-19. However, these approaches are becoming useless as the virus mutates. In this proposed research, the prediction model is trained by the proposed Correlated Feature Extraction (CFE) method according to the virus behaviors. Furthermore, two prediction models were established in this study. The first model uses initial symptoms to forecast positive cases. Second, based on the health status of the positive cases, the risk rate is estimated. In this study, the disease is forecast by Feed-Forward Artificial Neural Networks (FFANN). Finally, a comparative study has been conducted with the recently developed COVID-19 disease prediction methods to demonstrate the training efficacy and accuracy of the proposed COVID-19 forecasting system. © 2021 Seventh Sense Research Group®