Ceteris Paribus Profiles

Ceteris Paribus Profiles (What-If Plots) are designed to present model responses around selected points in a feature space. For example around a single prediction for an interesting observation. Plots are designed to work in a model-agnostic fashion, they are working for any predictive Machine Learning model and allow for model comparisons. Ceteris Paribus Plots supplement the Break Down Plots from 'breakDown' package.


News

ceterisParibus 0.3.1

  • The what_if_2d() function plots 2D ceteris paribus plots. This function returns an object of the class what_if_2d. One can use generic print() or plot() to show these profiles. Note that profiles for all pairs of variables are generated, thus it may be a time consuming operation if number of variables is high.
  • New function calculate_oscillations() that calculates variable importance measures for a single observation.

ceterisParibus 0.3

Major refactoring of the code

Please note, that what_if() and local_fit() functions are remainders from version 0.2 and they will be deprecated in version 0.4. They will not be avaliable in version 1.0. From version 0.3 the recommended way to create explainers is through the function ceteris_paribus().

Please note, that plot_interactive() function will be deprecated in version 0.4. It will not be avaliable in version 1.0.

Reference manual

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

install.packages("ceterisParibus")

0.3.1 by Przemyslaw Biecek, 3 months ago


https://pbiecek.github.io/ceterisParibus/


Report a bug at https://github.com/pbiecek/ceterisParibus/issues


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


Authors: Przemyslaw Biecek [aut, cre]


Documentation:   PDF Manual  


GPL-2 license


Imports DALEX, knitr

Depends on ggplot2, gower

Suggests randomForest, ggiraph, e1071, testthat, rpart


See at CRAN