lag_multiple_datasets: Create a spatio-temporal embed of multiple datasets

In mrfoliveira/STEvaluation-MDPI2021: Evaluation procedures for forecasting with spatio-temporal data

Description

Transform multiple spatio-temporal data sets in a list of lists into a list of lists containing data frames that have a spatio-temporal embed. Only data for "interior" points in the grid is kept. That is, points that do not have four immediate neighbours are filtered out.

Usage

lag_multiple_datasets(
  data_list,
  LAG_use,
  SLAGS,
  min_time = rep(max(LAG_use), length(LAG_use))
)

Arguments

data_list	A multi-level list as the ones generated by generate_multiple_datasets. The first level corresponds to a certain grid size; the second level to a certain time series size and the third level contains multiple lists containing datasets (that might have been generated from different STARMA specifications).
LAG_use	A vector of the temporal lag orders to use.
SLAGS	A list containing vectors of length of the corresponing temporal lag size. Each of the vectors contains values of 0 meaning no neighbour columns are used at that temporal lag, and/or values of 1 meaning that the values of immediate neighbours are included for that temporal lag.
min_time	A vector of the minimum time-stamps to be included in the datasets. Defaults to rep(max(LAG_use), length(LAG_use).

Value

A list of lists similar to data_list, but with the data sets having spatio-temporal embeds.

See Also

generate_multiple_datasets, st_lag_neib_ord1

mrfoliveira/STEvaluation-MDPI2021 documentation built on Dec. 21, 2021, 10:07 p.m.