Spatio-temporal point processes, partial likelihood, foot and mouth disease

Area of Science:

• Statistics
• Epidemiology
• Computational Science

Background:

• Spatio-temporal point process data are common in various scientific fields.
• Modeling these processes often involves complex conditional intensity functions.
• Traditional likelihood-based inference can be computationally intensive and difficult to implement.

Purpose of the Study:

• To develop a computationally straightforward and routinely applicable inference method for spatio-temporal point processes.
• To address the challenges posed by analytically intractable likelihoods in existing models.
• To demonstrate the utility of the proposed method on real-world epidemic data.

Main Methods:

• Proposed a partial likelihood approach as an alternative to full likelihood methods.
• The method bypasses the need for computationally intensive Monte Carlo simulations.
• Applied the partial likelihood method to a model of the 2001 UK foot and mouth disease epidemic.

Main Results:

• The partial likelihood method offers a computationally efficient alternative for spatio-temporal point process inference.
• The proposed method is straightforward to implement and can be applied routinely.
• Successful application to the foot and mouth epidemic data validates the approach.

Conclusions:

• Partial likelihood provides a practical and efficient solution for analyzing spatio-temporal point process data.
• This method reduces computational burden and complexity in modeling.
• The approach is suitable for epidemiological studies and other fields utilizing point process data.