lagData: Convert many-obs-per case data into lagged matrices

In atyre2/flmtools: Utilities for Working with Functional Linear Models

Description

lagData creates two matrix variables: one with the lagged covariates, and a second with the indices used in the smooth term from mgcv.

Usage

lagData(lagData, response, unit, startUnit, nUnits, measure)

Arguments

lagData	data.frame or tibble containing the predictor variables to be created as lags. Must be sorted by Location, year, month.
response	data.frame or tibble with the response and one-per-case covariates
unit	character string variable name in lagData by which to lag
startUnit	integer where to start the lagging
nUnits	integer how many rows to lag backwards
measure	character string name of column in lagData to convert to matrices

Details

Assumes that lagData have already been aggregated to the appropriate resolution. lagData should have nrows(response)*nUnits - startUnit *not true.

Value

returns a dataframe or tibble with as many rows as response, and two new matrix columns

Examples

test <- lagData(env_data, response_data, unit = "month", 6, 24, "envVar")
dim(test$envVar_6_24)
test$lag_6_24[1:5,1:5]

atyre2/flmtools documentation built on Oct. 21, 2020, 5:26 a.m.