hdp_addconparam: Add concentration parameters to a hdpState object

In nicolaroberts/hdp: Hierarchical Dirichlet Process for categorical count data

Description

Add concentration parameters to a hdpState object by specifying the shape and rate parameters of the gamma prior/s. DPs using these new concentration parameters can be added with hdp_adddp. Data is assigned via hdp_setdata. When initialised, the DP nodes are 'heldout' (not available for posterior sampling) and will need to be activated (see dp_activate). Finally, the posterior sampling process (a Gibbs sampler) is run via hdp_posterior.

Usage

hdp_addconparam(hdp, alphaa, alphab)

Arguments

hdp	A hdpState object
alphaa	Shape hyperparameters for the gamma priors over the DP concentration parameters.
alphab	Rate hyperparameters for the gamma priors over the DP concentration parameters.

Value

A hdpState object updated with the new concentration parameters. See hdpState-class

See Also

hdp_init, hdp_adddp, hdp_setdata, dp_activate, hdp_posterior

Examples

hdp_example <- hdp_init(c(0, 1, 1), c(1, 2, 2), rep(1, 6), rep(2, 2), rep(0.5, 2))
hdp_example <- hdp_addconparam(hdp_example, rep(1, 2), rep(1, 2))

nicolaroberts/hdp documentation built on May 23, 2019, 5:09 p.m.