- changed URLs in the DESCRIPTION as requested in (#484)
- Fix model_info documentation (#498)
- Support for yardstic metrics (#495)
- Changed default in
explain(colorize=) according to (#473)
- Added explain/yhat support for
explain() warns if target has more than two values for classification (#418)
model_performance_auc functions are rewritten to handle repeated predictions (#442)
plot function works for list of explanations (if possible) (#424)
- Order of explainer labels in different plots is the same. To get to this point, orders in
plot.model_performance(..., geom = "histogram" & "boxplot") are reversed (#400)
- Fixed multiclass explainer when data has one column (#405)
- Now explainer handles R functions (#396)
predict_parts function handles the
N argument natively (#394)
- All encouters of
nieghbour(s) (EN-spelling) were replaced with
neighbor(s) (US-spelling) for the consistency and backword compatibility.
- Fixed bug when
predict_diagnostics raised error if
neighbor value was higer than
- Added new parameter (
explain function that allows specifying positive class in binary classification tasks (#250).
model_diagnostics() returning an error when
- Fixed R package not working with Python Explainer (#318)
model_diagnostics() returning an error when
residuals is of
array class (#319)
- Fixed grid lines in
theme_drwhy on Windows
- Fixed logical values in y rising unnecessery warnings for classification task (#336)
plot.predict_diagnostics now passess ellipsis to
- This version requires
iBreakDown v1.3.1 and
plot.model_profile for multiple profiles (#237).
- External tests for not suggested packages added to gh-actions (#237).
- Extended and refreshed documentation (#237).
- All dontrun statements changed to donttest according to CRAN policy.
- Added value for
s parameter in
model_diagnostics passing wrong arguments to residual_function.
- Fixed aesthetic for
hist geometry in
plot.model_performance using wrong arugments.
model_performance will not work if
- Corrected description of
- New warning messages for
y parameter in
- Solved bug in
predicts_parts not to plot correctly when task is multiclass.
variable_effect is now deprecated
- fixed typo in
- rewrite tests
predict_parts class to objects and
model_parts class to objects and
- plot parameters added to the documentation
- Now in the
predict_profile function one can specify how grid points shall be calculated, see
predict_part function has two new options for type:
plot.model_performance function has a new
geom="prc" for Precision Recall curve (#273).
DALEX now fully supports multiclass classification.
explain() will use new residual function (1 - true class probability) if multiclass classification is detected.
model_performance() now support measures for multiclass classification.
ggpubr from suggests.
lossFunction argument is now deprecated in
plot.model_performance(). Use the
model_profile color changed to
colors_discrete_drwhy(1) which impacts the color of the line in
loss_name attribute added to loss functions. It will be passed to plot function for objects created with
- fixed tests and WARNINGs on CRAN
model_profile for Accumulated Local rofiles by default use centering (
center = TRUE)
n_sample argument in
model_parts (now it’s
iBreakDown are now imported by DALEX
- updated title for
explain removed check related to duplicated target variable (#164).
feature_response moved to
variable_effect and now it calls
prediction_breakdown moved to
variable_attribution and now it calls
variable_importance, not it calls the
lrm models from
theme_drwhy has now left aligned title and subtitle.
residuals_distribution calculates now diagnostic plots based on residuals (#143).
model_performance calculates several metrics for classification and regression models (#146).
plot.model_performance now supports ROC charts, LIFT charts, Cummulative Gain charts, histograms, boxplots and ecdf
residuals_distributon is now
individual_diagnostics and produces objects of the class
plot.individual_diagnostics_explainers now plots objects of the class
yhat for caret models now returns matrix instead of data.frame
model_diagnostics new function that plots residuals againes selected variable
- names of functions are changed to be compliant with latest version of the XAI pyramide
weights to the explainer. Note that not all explanations know how to handle weights (#118).
model_info() now support models created with
- new dataset
titanic_imputed as requested in (#104).
explain() function now detects if target variable
y is present in the
data as requested in (#103).
- the DALEX GitHub repository is transfered from
pbiecek/DALEX to ModelOriented/DALEX.
- Examples updated. Now they use only datasets available from DALEX.
- yhat.H2ORegressionModel and yhat.H2OBinomialModel moved to (DALEXtra) and merged into explain_h2o() function.
- yhat.WrappedModelmoved to (DALEXtra) and merged as explain_mlr() function.
- Wrapper for scikit-learn models restored in (DALEXtra) package.
- loss_one_minus_auc function added to loss_functions.R. It uses 1-auc to compute loss. Function created by Alicja Gosiewska.
- Extension for DALEX avaiable at (DALEXtra)
explain() function is more verbose. With
verbose = TRUE (default) it prints detailed information about elements of an explainer (#95).
- New support for scikit-learn models via
yhat functions for
caret packages (added by Szymon).
plot.variable_importance_explainer() has now
desc_sorting argument. If FALSE then variable importance will be sorted in an increasing order (#41).
ALEPlot are going to
Suggested. (#60). In next releases they will be deprecated.
predict function that calls the
predict_function hidden in the
explainer object. (#58).
titanic dataset is copied from
stablelearner package. Some features are transformed (some
NA replaced with
0, more numeric features).
DALEX is being prepared for tighter integration with
- temporally there is a duplicated
- Added new
- New arguments in the
bar_width with widths of bars and
show_baseline if baseline shall be included in these plots.
- New skin in the
- New skin in the
- Test datasets are now named
- For binary classification we return just a second column. NOTE: this may cause some unexpected problems with code dependend on defaults for DALEX 0.2.6.
- New versions of
- Residual distribution plots for model performance are now more legible when multiple models are plotted. The styling of plot and axis titles have also been improved (@kevinykuo).
- The defaults of
single_prediction() are now consistent with
baseline is now
0 by default instead of
"Intercept". The user can also specify the
baseline and other arguments by passing them to
single_prediction (@kevinykuo, #39). WARNING: Change in the default value of
yhat.* functions help to handle additional parameters to different
- New dataset
HRTest. Target variable is a factor with three levels. Is used in examples for classification.
plot.model_performance() has now
show_outliers parameter. Set it to anything >0 and observations with largest residuals will be presented in the plot. (#34)
- Small fixes in
variable_response() to better support of
gbm models (c8393120ffb05e2f3c70b0143c4e92dc91f6c823).
- Better title for
- Tested with
breakDown v 0.1.6.
single_variable() / variable_response() function uses
- New names for some functions:
prediction_breakdown(). Old names are now deprecated but still working. (#12)
- A new dataset
apartments - will be used in examples
variable_importance() allows work on full dataset if
n_sample is negative
plot_model_performance() uses ecdf or boxplots (depending on
single_variable() supports factor variables as well (with the use of
factorMerger package). Remember to use
type='factor' when playing with factors. (#10)
- Change in the function
explain(). Old version has an argument
predict.function, now it’s
predict_function. New name is more consistent with other arguments. (#7)
- New vigniette for
xgboost model (#11)
- Support for global model structure explainers with
- DALEX package is now public
explain() function implemented
single_prediction() function implemented
single_variable() function implemented