clear_outcome: Partially clear outcome variable in new data by overriding...

In krzjoa/torchts: Time series Models with torch

Partially clear outcome variable in new data by overriding with NA values

Description

Partially clear outcome variable in new data by overriding with NA values

Usage

clear_outcome(data, index, outcome, timesteps, key = NULL)

Arguments

data	(data.frame) New data
index	Date variable
outcome	Outcome (target) variable
timesteps	(integer) Number of timesteps used by RNN model
key	A key (id) to group the data.frame (for panel data)

Examples

tarnow_temp <-
  weather_pl %>%
  filter(station == "TRN") %>%
  select(date, tmax_daily, tmin_daily, press_mean_daily)

TIMESTEPS <- 20
HORIZON   <- 1

data_split <-
  time_series_split(
    tarnow_temp, date,
    initial = "18 years",
    assess  = "2 years",
    lag     = TIMESTEPS
  )

cleared_new_data <-
  testing(data_split) %>%
  clear_outcome(date, tmax_daily, TIMESTEPS)

head(cleared_new_data, TIMESTEPS + 10)

krzjoa/torchts documentation built on June 24, 2022, 5:30 a.m.