Quickly Score Models in Data Science and Machine Learning

A set of tools for quickly scoring models in data science and machine learning. This toolset is written in C++ for blazing fast performance.

Project Status: Active - The project has reached a stable, usable state and is being actively developed. R Version 3.2.3+ CRAN_Status_Badge Downloads from the RStudio CRAN mirror Build Status Build status codecov.io

scorer is a set of tools for quickly scoring models in data science and machine learning. This toolset is written in C++, where possible, for blazing fast performance. This toolset's API follows that of sklearn.metrics as closely as possible so one can easily switch back and forth between the two languages without too much cognitive dissonance. The following types of metrics are currently implemented in scorer:

  • Regression metrics (implemented in 0.2.0)

The following types of metrics are soon to be implemented in scorer:

  • Classification metrics (to be implemented in 0.3.0)
  • Multilabel ranking metrics (to be implemented in 0.3.0)
  • Clustering metrics (to be implemented in 0.3.0)
  • Biclustering metrics (to be implemented in 0.3.0)
  • Pairwise metrics (to be implemented in 0.3.0)


You can install:

  • the latest released version from CRAN with


  • the latest development version from Github with


    if (packageVersion("devtools") < 1.6) {

If you encounter a clear bug, please file a minimal reproducible example on github.



  • All functions from scorer 0.1.0 have been deprecated in favor of a new API that mirrors the API of sklearn.metrics. These functions will be removed in 1.0.0.
  • Added more functions!
  • Nearly all functions implemented in C++ for blazing fast speed!
  • Additional features such as sample weighting for some error metrics have been identified and placed on a backburner for future releases.
  • Implemented unit tests for base functions.

scorer 0.1.0


  • Implemented several functions for estimating errors.
  • Implemented unit tests for nearly all functions.
  • First minor version release to CRAN!

Bug fixes

  • Fixed minor error in passing multiple arguments to mae().


Regression metrics

Load library and data

#> [1] '0.2.0'

Visualize data

ggplot(mtcars, aes(x = wt, y = mpg)) + 
  geom_point() + 
  geom_smooth(method = 'lm') + 
  expand_limits(x = c(0, 6), y = c(0, 40))

Partition data into train and test sets

n_train <- floor(nrow(mtcars) * 0.60)
n_test <- nrow(mtcars) - n_train
mask <- sample(c(rep(x = TRUE, times = n_train), rep(x = FALSE, times = n_test)))
mtcars[, "Type"] <- ifelse(mask, "Train", "Test")
train_mtcars <- mtcars[mask, ]
test_mtcars <- mtcars[!mask, ]
ggplot(mtcars, aes(x = wt, y = mpg, color = Type)) + 
  geom_point() + 
  expand_limits(x = c(0, 6), y = c(0, 40))

Build a model on train data set

model <- lm(mpg ~ wt, data = train_mtcars)

Predict model using the test data set

test_mtcars[, "predicted_mpg"] <- predict(model, newdata = test_mtcars)

Score model using various metrics

scorer::mean_absolute_error(test_mtcars[, "mpg"], test_mtcars[, "predicted_mpg"])
#> [1] 3.287805
scorer::mean_squared_error(test_mtcars[, "mpg"], test_mtcars[, "predicted_mpg"])
#> [1] 15.43932

Build a final model on all the data

final_model <- lm(mpg ~ wt, data = mtcars)

Predict final model using the original data set

mtcars[, "predicted_mpg"] <- predict(final_model, newdata = mtcars)

Score final model using various metrics

scorer::explained_variance_score(mtcars[, "mpg"], mtcars[, "predicted_mpg"])
#> [1] 847.7252
scorer::unexplained_variance_score(mtcars[, "mpg"], mtcars[, "predicted_mpg"])
#> [1] 278.3219
scorer::total_variance_score(mtcars[, "mpg"], mtcars[, "predicted_mpg"])
#> [1] 1126.047
scorer::r2_score(mtcars[, "mpg"], mtcars[, "predicted_mpg"])
#> [1] 0.7528328

Classification metrics





Session Information

#> R version 3.2.3 (2015-12-10)
#> Platform: x86_64-apple-darwin13.4.0 (64-bit)
#> Running under: OS X 10.11.3 (El Capitan)
#> locale:
#> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> other attached packages:
#> [1] ggplot2_2.0.0 scorer_0.2.0 
#> loaded via a namespace (and not attached):
#>  [1] Rcpp_0.12.3      digest_0.6.9     plyr_1.8.3       grid_3.2.3      
#>  [5] gtable_0.1.2     formatR_1.2.1    magrittr_1.5     evaluate_0.8    
#>  [9] scales_0.3.0     stringi_1.0-1    rmarkdown_0.8.1  labeling_0.3    
#> [13] tools_3.2.3      stringr_1.0.0    munsell_0.4.2    yaml_2.1.13     
#> [17] colorspace_1.2-6 htmltools_0.2.6  knitr_1.12


Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.2.0 by Paul Hendricks, 3 years ago


Report a bug at https://github.com/paulhendricks/scorer/issues

Browse source code at https://github.com/cran/scorer

Authors: Paul Hendricks [aut, cre]

Documentation:   PDF Manual  

MIT + file LICENSE license

Imports Rcpp

Suggests testthat

Linking to Rcpp

See at CRAN