mcmc.DrawFutureTransactions: Draws number of future transactions based on MCMC parameter...

In BTYDplus: Probabilistic Models for Assessing and Predicting your Customer Base

Description

For each customer and each provided MCMC parameter draw this method will sample the number of transactions during the holdout period T.star. If argument size is provided then it returns a flexible number of draws, whereas for each customer and each draw it will first make a draw from the parameter draws.

Usage

mcmc.DrawFutureTransactions(
  cal.cbs,
  draws,
  T.star = cal.cbs$T.star,
  sample_size = NULL
)

Arguments

cal.cbs	Calibration period customer-by-sufficient-statistic (CBS) data.frame.
draws	MCMC draws as returned by *.mcmc.DrawParameters
T.star	Length of period for which future transactions are counted.
sample_size	Number of samples to draw. Defaults to the same number of parameter draws that are passed to draws.

Value

2-dim matrix [draw x customer] with sampled future transactions.

Examples

data("groceryElog")
cbs <- elog2cbs(groceryElog, T.cal = "2006-12-31")
param.draws <- pnbd.mcmc.DrawParameters(cbs,
  mcmc = 100, burnin = 50, thin = 10, chains = 1) # short MCMC to run demo fast
xstar.draws <- mcmc.DrawFutureTransactions(cbs, param.draws)
cbs$xstar.est <- apply(xstar.draws, 2, mean)
cbs$pactive <- mcmc.PActive(xstar.draws)
head(cbs)

BTYDplus documentation built on Jan. 21, 2021, 5:10 p.m.