load.Hst.ls.2Zs: Load Observations into Space-Time Covariates

Description

Insert an observation matrix into space-time covariates, but segregate based on missing values

Usage

load.Hst.ls.2Zs(Z, Z.na, Hst.ls.Z, xwhich, rgr.lags = c(0))

Value

An unnamed list of length \(\tau\), each element will be a numeric \(n\) x \(p_st\) matrix.

Arguments

Z: Observation data. A \(\tau\) x \(n\) numeric matrix.
Z.na: Missing data indicator. A \(\tau\) x \(n\) boolean matrix.
Hst.ls.Z: Space-time covariates. A list of length \(\tau\), each element should be a numeric \(n\) x \(p_st\) matrix.
xwhich: Which column-pair of Hst.ls.Z[[i]] to insert into the \(i\)th row of Z. A scalar positive integer. By 'column-pair', we mean, e.g., a value of 1 will fill columns 1 and 2, a value of 2 will fill columns 3 and 4, a value of 3 will fill columns 5 and 6, etc.
rgr.lags: Temporal lagging of Z. A scalar integer.

Details

This function, along with load.Hst.ls.Z, allows the user to convert a set of observations into covariates for another set of observations. Unlike load.Hst.ls.Z, this function splits Z based on the argument Z.na. Values associated with FALSE elements of Z.na are placed into the first column of the specified column-pair of Hst.ls.Z, Values associated with TRUE elements of Z.na are placed into the second column of the specified column-pair of Hst.ls.Z (all other values in in the specified column-pair of Hst.ls.Z are zeroed).

Examples

Run this code



###### here's an itty-bitty example

tau <- 7
n <- 5

Z <- matrix(1, tau, n)

Z.na <- matrix(FALSE, tau, n)
Z.na[2:3, 4] <- TRUE

Z[Z.na] <- 2

Hst.ls <- list()
for(i in 1:tau) { Hst.ls[[i]] <- matrix(rnorm(n*4), nrow=n) }


load.Hst.ls.2Zs(Z, Z.na, Hst.ls.Z=Hst.ls, 1, 0)


########## insert into cols 3 and 4

load.Hst.ls.2Zs(Z, Z.na, Hst.ls.Z=Hst.ls, 2, 0)