Genetic algorithm with cross-validation-based epidemic model and application to the early diffusion of COVID-19 in...

Area of Science:

• Epidemiology
• Mathematical Modeling
• Computational Biology

Background:

• Dynamical epidemic models are crucial for understanding disease spread.
• Overfitting can compromise the accuracy of these models.
• Genetic algorithms and cross-validation offer potential solutions for model optimization.

Purpose of the Study:

• To develop and validate an optimized dynamical epidemic model using a genetic algorithm and cross-validation.
• To apply the model to study COVID-19 outbreak dynamics in Algeria.
• To estimate key epidemic parameters, including the asymptomatic infected population.

Main Methods:

• A dynamical compartmental model was optimized using a genetic algorithm.
• Cross-validation was employed to prevent overfitting, splitting data into training and validation subsets.
• The model was initially tested on data from Italy, Spain, Germany, and South Korea before application to Algeria.

Main Results:

• An inverse relationship was observed between training sample size and genetic algorithm generations.
• The model accurately estimated epidemic parameters and the asymptomatic infected proportion in Algeria.
• Key metrics such as the basic reproduction number (3.78) and effective reproduction number (0.651) were calculated for Algeria.

Conclusions:

• The enhanced compartmental model is a reliable tool for epidemic forecasting and parameter estimation.
• The study provides valuable insights into the COVID-19 outbreak dynamics in Algeria.
• Publicly accessible code facilitates reproducibility and further research in epidemic modeling.