Build a Time Series Recipe — ts_auto

Automatically builds generic time series recipe objects from a given tibble.

Usage

ts_auto_recipe(
  .data,
  .date_col,
  .pred_col,
  .step_ts_sig = TRUE,
  .step_ts_rm_misc = TRUE,
  .step_ts_dummy = TRUE,
  .step_ts_fourier = TRUE,
  .step_ts_fourier_period = 365/12,
  .K = 1,
  .step_ts_yeo = TRUE,
  .step_ts_nzv = TRUE
)

Arguments

.data

The data that is going to be modeled. You must supply a tibble.

.date_col

The column that holds the date for the time series.

.pred_col

The column that is to be predicted.

.step_ts_sig

A Boolean indicating should the timetk::step_timeseries_signature() be added, default is TRUE.

.step_ts_rm_misc

A Boolean indicating should the following items be removed from the time series signature, default is TRUE.

iso$
xts$
hour
min
sec
am.pm

.step_ts_dummy

A Boolean indicating if all_nominal_predictors() should be dummied and with one hot encoding.

.step_ts_fourier

A Boolean indicating if timetk::step_fourier() should be added to the recipe.

.step_ts_fourier_period

A number such as 365/12, 365/4 or 365 indicting the period of the fourier term. The numeric period for the oscillation frequency.

.K

The number of orders to include for each sine/cosine fourier series. More orders increase the number of fourier terms and therefore the variance of the fitted model at the expense of bias. See details for examples of K specification.

.step_ts_yeo

A Boolean indicating if the recipes::step_YeoJohnson() should be added to the recipe.

.step_ts_nzv

A Boolean indicating if the recipes::step_nzv() should be run on all predictors.

Details

This will build out a couple of generic recipe objects and return those items in a list.

Author

Steven P. Sanderson II, MPH

Examples

suppressPackageStartupMessages(library(dplyr))
suppressPackageStartupMessages(library(rsample))

data_tbl <- ts_to_tbl(AirPassengers) %>%
  select(-index)

splits <- initial_time_split(
 data_tbl
 , prop = 0.8
)

ts_auto_recipe(
    .data = data_tbl
    , .date_col = date_col
    , .pred_col = value
)
#> $rec_base
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> $rec_date
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> ── Operations 
#> • Timeseries signature features from: ds
#> • Centering and scaling for: dplyr::contains("index.num"), ...
#> • Variables removed: dplyr::matches("(iso$)|(xts$)|(hour)|(min)|(sec)|(am.pm)")
#> • Dummy variables from: recipes::all_nominal_predictors()
#> 
#> $rec_date_fourier
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> ── Operations 
#> • Timeseries signature features from: ds
#> • Centering and scaling for: dplyr::contains("index.num"), ...
#> • Variables removed: dplyr::matches("(iso$)|(xts$)|(hour)|(min)|(sec)|(am.pm)")
#> • Dummy variables from: recipes::all_nominal_predictors()
#> • Fourier series features from: ds
#> • Yeo-Johnson transformation on: value
#> 
#> $rec_date_fourier_nzv
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> ── Operations 
#> • Timeseries signature features from: ds
#> • Centering and scaling for: dplyr::contains("index.num"), ...
#> • Variables removed: dplyr::matches("(iso$)|(xts$)|(hour)|(min)|(sec)|(am.pm)")
#> • Dummy variables from: recipes::all_nominal_predictors()
#> • Fourier series features from: ds
#> • Yeo-Johnson transformation on: value
#> • Sparse, unbalanced variable filter on: recipes::all_predictors()
#> 

ts_auto_recipe(
  .data = training(splits)
  , .date_col = date_col
  , .pred_col = value
)
#> $rec_base
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> $rec_date
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> ── Operations 
#> • Timeseries signature features from: ds
#> • Centering and scaling for: dplyr::contains("index.num"), ...
#> • Variables removed: dplyr::matches("(iso$)|(xts$)|(hour)|(min)|(sec)|(am.pm)")
#> • Dummy variables from: recipes::all_nominal_predictors()
#> 
#> $rec_date_fourier
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> ── Operations 
#> • Timeseries signature features from: ds
#> • Centering and scaling for: dplyr::contains("index.num"), ...
#> • Variables removed: dplyr::matches("(iso$)|(xts$)|(hour)|(min)|(sec)|(am.pm)")
#> • Dummy variables from: recipes::all_nominal_predictors()
#> • Fourier series features from: ds
#> • Yeo-Johnson transformation on: value
#> 
#> $rec_date_fourier_nzv
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs 
#> Number of variables by role
#> outcome:   1
#> predictor: 1
#> 
#> ── Operations 
#> • Timeseries signature features from: ds
#> • Centering and scaling for: dplyr::contains("index.num"), ...
#> • Variables removed: dplyr::matches("(iso$)|(xts$)|(hour)|(min)|(sec)|(am.pm)")
#> • Dummy variables from: recipes::all_nominal_predictors()
#> • Fourier series features from: ds
#> • Yeo-Johnson transformation on: value
#> • Sparse, unbalanced variable filter on: recipes::all_predictors()
#>