arma_benchmark {spatPomp}

Calculated log-ARMA log-likelihood benchmark for spatPomp models

Description

Fits independent log-ARMA models for each unit, and calculates the conditional log-likelihood for each observation, as well as log-likelihood for each unit and total log-likelihood. A simple tool, but one with practical applicability, as demonstrated by King et al (2008) and Wheeler et al (2023). This function is designed for non-negative data, and adds 1 to each observation to avoid log(0).

Usage

arma_benchmark(spo, order = c(2, 0, 1))

Arguments

spo

A spatPomp object

order

A triple (p,d,q) for the ARIMA model fitted to the data. It is intended that d=0

Author(s)

Edward L. Ionides

References

King, A. A., Ionides, E. L., Pascual, M. and Bouma, M. J. (2008). Inapparent infections and cholera dynamics. Nature 454 877-880.

Wheeler, J., Rosengart, A. L., Jiang, Z., Tan, K., Treutle, N. and Ionides, E. L. (2023). Informing policy via dynamic models: Cholera in Haiti. arxiv:2301.08979.

See Also

Other utilities: expand_params()

Examples

# Complete examples are provided in the package tests
## Not run: 
m <- he10(U = 5)
arma_benchmark(m)

## End(Not run)