Sequential Monte Carlo Squared for online inference in stochastic epidemic models.

Area of Science:

• Epidemiology
• Computational Statistics
• Mathematical Modeling

Background:

• Effective epidemic modeling and surveillance demand computationally efficient methods for continuous parameter updates.
• Real-time tracking requires methods that can adapt to new data rapidly.

Purpose of the Study:

• To explore the application of an online variant of Sequential Monte Carlo Squared (O-SMC2) for real-time epidemic tracking using the Susceptible-Exposed-Infectious-Removed (SEIR) model.
• To assess the computational efficiency and accuracy of O-SMC2 in estimating epidemiological parameters.

Main Methods:

• Utilized an online variant of Sequential Monte Carlo Squared (O-SMC2) with a particle Metropolis-Hastings kernel.
• Applied O-SMC2 to simulated epidemic data and a real-world COVID-19 dataset from Ireland.
• Focused on using a fixed window of recent observations for parameter updates.

Main Results:

• Demonstrated the computational efficiency of O-SMC2 on simulated data.
• Successfully tracked a COVID-19 epidemic and estimated a time-dependent reproduction number.
• Achieved accurate online estimates of static and dynamic epidemiological parameters with reduced computational cost.

Conclusions:

• O-SMC2 provides accurate online estimates for epidemiological parameters, enhancing real-time epidemic monitoring.
• The method's computational efficiency makes it suitable for adaptive public health interventions.
• O-SMC2 offers a significant improvement over standard SMC2 for time-sensitive epidemic analysis.