doAggr: Generates a sequence of Poisson multinomial random values for...

In bogdanoancea/aggregation: R package for aggregation of device numbers into population number

Description

Generates a sequence of Poisson multinomial random values for a sequence of time instants. This function is called internally by rNnetEvent function which partitions the time instants into equal chunks, creates a cluster of working processes and distributes the chunks to different workers that run in parallel. This is an internal function that cannot be directly called by the users.

Usage

doAggr(ichunks, n, tiles, postLoc, dupProbs, regions)

Arguments

ichunks	A partition of time instants.
n	The number of random values to be generated.
tiles	The list of tile indexes.
postLoc	A list with posterior location probabilities matrices. Each element of the list corresponds to a device.
dupProbs	The duplicity probabilities for each device. It is a data.table object with two columns: deviceID, dupProb.
regions	a data.table object with two columns: tile, region. It defines each region as a composition of multiple tiles.

Value

a data.table object with the following column: time, region, N. It contains n random values for each combination time-region. An estimation of the number of individuals in a region can be obtained using the mean, mode or median of these values.

bogdanoancea/aggregation documentation built on Nov. 29, 2020, 5:09 p.m.