Adaptive best subset selection algorithm and genetic algorithm aided ensemble learning method identified a robust...

Area of Science:

• Medical Informatics
• Computational Biology
• Epidemiology

Background:

• Predicting COVID-19 patient severity is crucial for resource allocation and treatment strategies.
• Existing models may lack generalizability across diverse patient populations.
• The need for robust and stable predictive tools for infectious diseases is paramount.

Purpose of the Study:

• To develop and validate a stable prediction model for Coronavirus Disease 2019 (COVID-19) patient severity.
• To assess the model's performance in multicenter settings.
• To provide a reliable tool for clinical decision support in managing COVID-19.

Main Methods:

• An integrated ensemble learning approach was employed.
• The prediction model was constructed using diverse datasets.
• Validation was performed on independent multicenter patient cohorts.

Main Results:

• The developed ensemble learning model demonstrated high stability.
• The prediction model showed consistent performance across different healthcare settings.
• The model effectively predicted COVID-19 patient severity.

Conclusions:

• Integrated ensemble learning offers a robust method for building stable disease severity prediction models.
• The validated model can aid clinicians in assessing COVID-19 patient outcomes.
• Multicenter validation ensures the generalizability and reliability of the predictive tool for public health applications.