Hybrid optimized feature selection and deep learning based COVID-19 disease prediction.

Area of Science:

• Computational biology and bioinformatics
• Infectious disease modeling
• Machine learning applications in healthcare

Background:

• The COVID-19 pandemic has caused significant global mortality and necessitated widespread lockdowns.
• Current COVID-19 detection methods, such as RT-PCR, lack optimal sensitivity and effectiveness.
• There is a critical need for advanced diagnostic tools to improve COVID-19 detection accuracy.

Purpose of the Study:

• To propose an efficient and sensitive scheme for COVID-19 detection using artificial intelligence.
• To enhance the accuracy of COVID-19 case detection and prediction.
• To introduce a novel hybrid optimization and deep learning model for real-time analysis.

Main Methods:

• Utilized COVID-19 case data for detection and analysis.
• Employed Mayfly with Fruit Fly Optimization (MFFO) for feature selection.
• Developed a Conditional Autoregressive Value at Risk MFFO (Caviar-MFFO) to train a Deep Long Short-Term Memory (Deep LSTM) network for detection.

Main Results:

• The Caviar-MFFO-assisted Deep LSTM model demonstrated efficient performance in detecting COVID-19 cases.
• Achieved minimal Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) for recovered cases (1.438 and 1.199, respectively).
• The model reported MSE and RMSE values of 4.582 and 2.140 for death cases, and 6.127 and 2.475 for infected cases.

Conclusions:

• The proposed Caviar-MFFO-assisted Deep LSTM scheme offers a highly effective approach for COVID-19 detection.
• The model shows significant potential in improving the accuracy of epidemiological analysis and case management.
• This AI-driven method provides a promising alternative to conventional diagnostic techniques for infectious diseases.