characterize

The goal of characterize is to report data driven characteristics of a model fit that describe the model. For example:

the number of predictors used,
the number of terminal nodes,

and so on. These are interesting characteristics of the model that often reflect complexity of the model.

Installation

You can install the development version of characterize like so:

require(remotes)
remotes::install_github("topepo/characterize")

Example

This is a basic example which shows you how to solve a common problem:

library(characterize)
library(rpart)

cart_mod <- rpart(mpg ~ ., data = mtcars)
cart_mod
#> n= 32 
#> 
#> node), split, n, deviance, yval
#>       * denotes terminal node
#> 
#> 1) root 32 1126.04700 20.09062  
#>   2) cyl>=5 21  198.47240 16.64762  
#>     4) hp>=192.5 7   28.82857 13.41429 *
#>     5) hp< 192.5 14   59.87214 18.26429 *
#>   3) cyl< 5 11  203.38550 26.66364 *

characterize(cart_mod)
#> # A tibble: 2 × 3
#>   .metric             .estimator .estimate
#>   <chr>               <chr>          <dbl>
#> 1 num_active_features model              2
#> 2 num_term_nodes      model              3

# Also some low-level functions: 
.pluck_active_features(cart_mod) %>% 
  tidyr::unnest(value)
#> # A tibble: 2 × 2
#>   statistic       value
#>   <chr>           <chr>
#> 1 active_features cyl  
#> 2 active_features hp

Objects with at least one method: bagger, bart, brulee_mlp, C5.0, cforest, cubist, earth, glmnet, ksvm, lgb.Booster, mixo_pls, mixo_plsda, mixo_spls, mixo_splsda, multinom, nnet, party, partynode, ranger, rpart, terms, xgb.Booster, xrf.

Code of Conduct

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