Mix: Specify a Mix prior.

In StatisticsNZ/demest: Bayesian Demographic Estimation and Forecasting

Specify a Mix prior.

Description

A Mix prior is used to model an iteraction that includes an "along" dimension (typically, but not necessarily, time), and additional dimensions. Values for the additional dimensions evolve, generally smoothly, across successive values of the "along" dimension. A typical application is modelling age-structures that change gradually over time.

Usage

Mix(
  along = NULL,
  components = Components(),
  weights = Weights(),
  covariates = NULL,
  error = Error(),
  maxComponents = 10
)

Arguments

along	Name of a dimension.
components	An object of class Components.
weights	An object of class Weights.
covariates	An object of class Covariates or NULL (the default). If covariates is NULL, the prior does not include covariates.
error	An object of class Error.
maxComponents	The maximum number of components. Defaults to 10.

Details

The prior is modified from a model developed by Kunihama and Dunson (2013). The model is non-parametric, meaning that it extracts structure from the data with minimal guidance from the user. It is designed for use with sparse data.

References

Kunihama, T and Dunson, DB. 2013. Bayesian modeling of temporal dependence in large sparse contingency tables. Journal of the American Statistical Association, 108(504): 1324-1338.

See Also

Priors for the components can be specified using function Components, and weights (ie mixing parameters) can be specified using function Weights. However, in normal usage, it is typically best to stick with the defaults.

Examples

Mix()

StatisticsNZ/demest documentation built on Nov. 2, 2023, 7:56 p.m.