Biserial targeted feature projection based radial kernel regressive deep belief neural learning for covid-19...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Computational Biology

Background:

• Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, necessitates accurate and timely prediction methods.
• Existing prediction techniques for COVID-19 often lack accuracy and are time-consuming.
• There is a need for advanced computational models to improve diagnostic efficiency.

Purpose of the Study:

• To introduce a novel technique, Biserial Targeted Feature Projection-based Radial Kernel Regressive Deep Belief Neural Learning (BTFP-RKRDBNL), for enhanced COVID-19 prediction.
• To improve the accuracy and reduce the time consumption of COVID-19 disease prediction.
• To achieve higher accuracy and a lower false positive rate in disease identification.

Main Methods:

• The BTFP-RKRDBNL technique utilizes deep belief neural learning with multiple layers (two visible, two hidden).
• Point Biserial Correlative Target feature projection is employed to select relevant features and discard irrelevant ones, optimizing prediction time.
• Radial Kernel Regression analyzes training and testing features to identify COVID-19 presence.

Main Results:

• The BTFP-RKRDBNL method demonstrated significant improvements in prediction metrics.
• Prediction accuracy increased by 10%, sensitivity by 6%, and specificity by 21% compared to existing methods.
• Prediction time was reduced by 10%, indicating enhanced efficiency.

Conclusions:

• The BTFP-RKRDBNL technique offers a more accurate and efficient approach to COVID-19 prediction.
• This novel method addresses limitations of current prediction techniques by improving accuracy and reducing diagnostic time.
• The findings suggest potential for widespread clinical application in infectious disease diagnostics.