Plot Ceteris Paribus 2D Explanations

Source: R/plot_ceteris_paribus_2d.R

This function plots What-If Plots for a single prediction / observation.

# S3 method for ceteris_paribus_2d_explainer
plot(
  x,
  ...,
  facet_ncol = NULL,
  add_raster = TRUE,
  add_contour = TRUE,
  bins = 3,
  add_observation = TRUE,
  pch = "+",
  size = 6
)

Arguments

x

a ceteris paribus explainer produced with the ceteris_paribus_2d() function

...

currently will be ignored

facet_ncol

number of columns for the facet_wrap

add_raster

if TRUE then geom_raster will be added to present levels with diverging colors

add_contour

if TRUE then geom_contour will be added to present contours

bins

number of contours to be added

add_observation

if TRUE then geom_point will be added to present observation that is explained

pch

character, symbol used to plot observations

size

numeric, size of individual datapoints

Value

a ggplot2 object

References

Explanatory Model Analysis. Explore, Explain, and Examine Predictive Models. https://ema.drwhy.ai/

Examples

library("DALEX")
library("ingredients")
library("ranger")

# \donttest{
apartments_rf_model <- ranger(m2.price ~., data = apartments)

explainer_rf <- explain(apartments_rf_model,
                        data = apartments_test[,-1],
                        y = apartments_test[,1],
                        verbose = FALSE)

new_apartment <- apartments_test[1,]
new_apartment
#>      m2.price construction.year surface floor no.rooms    district
#> 1001     4644              1976     131     3        5 Srodmiescie

wi_rf_2d <- ceteris_paribus_2d(explainer_rf, observation = new_apartment)
head(wi_rf_2d)
#>      y_hat new_x1 new_x2            vname1  vname2  label
#> 1 5222.393   1920   20.0 construction.year surface ranger
#> 2 5227.252   1920   21.3 construction.year surface ranger
#> 3 5247.392   1920   22.6 construction.year surface ranger
#> 4 5232.506   1920   23.9 construction.year surface ranger
#> 5 5242.054   1920   25.2 construction.year surface ranger
#> 6 5237.366   1920   26.5 construction.year surface ranger

plot(wi_rf_2d)
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.

plot(wi_rf_2d, add_contour = FALSE)
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.

plot(wi_rf_2d, add_observation = FALSE)
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.

plot(wi_rf_2d, add_raster = FALSE)
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?


# HR data
model <- ranger(status ~ gender + age + hours + evaluation + salary, data = HR,
                probability = TRUE)

pred1 <- function(m, x)   predict(m, x)$predictions[,1]

explainer_rf_fired <- explain(model,
                              data = HR[,1:5],
                              y = as.numeric(HR$status == "fired"),
                              predict_function = pred1,
                              label = "fired")
#> Preparation of a new explainer is initiated
#>   -> model label       :  fired 
#>   -> data              :  7847  rows  5  cols 
#>   -> target variable   :  7847  values 
#>   -> predict function  :  pred1 
#>   -> predicted values  :  No value for predict function target column. (  default  )
#>   -> model_info        :  package ranger , ver. 0.14.1 , task classification (  default  ) 
#>   -> predicted values  :  numerical, min =  0.0003492977 , mean =  0.3636236 , max =  0.9849971  
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -0.7439001 , mean =  0.0002096664 , max =  0.8139948  
#>   A new explainer has been created!  
new_emp <- HR[1,]
new_emp
#>   gender      age    hours evaluation salary status
#> 1   male 32.58267 41.88626          3      1  fired

wi_rf_2d <- ceteris_paribus_2d(explainer_rf_fired, observation = new_emp)
head(wi_rf_2d)
#>       y_hat   new_x1   new_x2 vname1 vname2 label
#> 1 0.4340871 20.00389 35.00000    age  hours fired
#> 2 0.4958088 20.00389 35.44978    age  hours fired
#> 3 0.5366813 20.00389 35.89955    age  hours fired
#> 4 0.5343380 20.00389 36.34933    age  hours fired
#> 5 0.5209654 20.00389 36.79911    age  hours fired
#> 6 0.5449915 20.00389 37.24889    age  hours fired

plot(wi_rf_2d)
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: The following aesthetics were dropped during statistical transformation: fill
#> ℹ This can happen when ggplot fails to infer the correct grouping structure in
#>   the data.
#> ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
#>   variable into a factor?
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.
#> Warning: Raster pixels are placed at uneven horizontal intervals and will be shifted
#> ℹ Consider using `geom_tile()` instead.

# }