Tidy Modelling for Nested Data.

nestedmodels

The goal of nestedmodels is to allow the modelling of nested data. Some models only accept certain predictors. For panel data, it is often desirable to create a model for each panel. nestedmodels enhances the ‘tidymodels’ set of packages by allowing the user to classify a model as ‘nested’.

Installation

# Install the released version on CRAN
install.packages("nestedmodels")

# Or install the development version from GitHub:
# install.packages("devtools")
devtools::install_github("ashbythorpe/nestedmodels")

Example

library(nestedmodels)

Nested models are often best used on panel data.

data <- example_nested_data

nested_data <- tidyr::nest(example_nested_data, data = -id)

nested_data
#> # A tibble: 20 × 2
#>       id data             
#>    <int> <list>           
#>  1     1 <tibble [50 × 6]>
#>  2     2 <tibble [50 × 6]>
#>  3     3 <tibble [50 × 6]>
#>  4     4 <tibble [50 × 6]>
#>  5     5 <tibble [50 × 6]>
#>  6     6 <tibble [50 × 6]>
#>  7     7 <tibble [50 × 6]>
#>  8     8 <tibble [50 × 6]>
#>  9     9 <tibble [50 × 6]>
#> 10    10 <tibble [50 × 6]>
#> 11    11 <tibble [50 × 6]>
#> 12    12 <tibble [50 × 6]>
#> 13    13 <tibble [50 × 6]>
#> 14    14 <tibble [50 × 6]>
#> 15    15 <tibble [50 × 6]>
#> 16    16 <tibble [50 × 6]>
#> 17    17 <tibble [50 × 6]>
#> 18    18 <tibble [50 × 6]>
#> 19    19 <tibble [50 × 6]>
#> 20    20 <tibble [50 × 6]>

The nested_resamples() function makes sure that the testing and training data contain every unique value of ‘id’.

split <- nested_resamples(nested_data, rsample::initial_split())

data_tr <- rsample::training(split)
data_tst <- rsample::testing(split)

Fitting a nested model to this data is very simple.

model <- parsnip::linear_reg() %>%
  nested()

fit <- fit(model, z ~ x + y + a + b, 
           tidyr::nest(data_tr, data = -id))

predict(fit, data_tst)
#> # A tibble: 260 × 1
#>    .pred
#>    <dbl>
#>  1  35.0
#>  2  27.7
#>  3  35.0
#>  4  39.4
#>  5  30.4
#>  6  29.5
#>  7  33.8
#>  8  33.1
#>  9  26.3
#> 10  18.9
#> # ℹ 250 more rows

If you don’t want to nest your data manually, use step_nest() inside a workflow:

recipe <- recipes::recipe(data_tr, z ~ x + y + a + b + id) %>%
  step_nest(id)

wf <- workflows::workflow() %>%
  workflows::add_model(model) %>%
  workflows::add_recipe(recipe)

wf_fit <- fit(wf, data_tr)

predict(wf_fit, data_tst)
#> # A tibble: 260 × 1
#>    .pred
#>    <dbl>
#>  1  35.0
#>  2  27.7
#>  3  35.0
#>  4  39.4
#>  5  30.4
#>  6  29.5
#>  7  33.8
#>  8  33.1
#>  9  26.3
#> 10  18.9
#> # ℹ 250 more rows

Please note that the nestedmodels project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.