stochastic loglikelihood constructor good for large datasets. if applied to a gradient ascent method, this will perform stochastic gradient ascent.

stochastic loglikelihood constructor good for large datasets. if applied to a gradient ascent method, this will perform stochastic gradient ascent.

Usage

stochastic_loglike(log.p, obs, options)

Arguments

log.p

log pdf (or pmf) of the parametric model being fit to `obs` parameters. it can also just be proportional to the log pdf, since sometimes the normalizing constant is unknown or hard to compute.

obs

a matrix, vector, or data frame of observations

options

a list of options