plot.variable_dropout_explainer: Plots Global Model Explanations (Variable Drop-out)

Description

Function plot.variable_dropout_explainer plots dropouts for variables used in the model. It uses output from variable_dropout function that corresponds to permutation based measure of variable importance.

Usage

# S3 method for variable_dropout_explainer
plot(x, ..., max_vars = 10)

Arguments

a variable fropout exlainer produced with the 'variable_dropout' function

...

other explainers that shall be plotted together

max_vars

maximum number of variables that shall be presented for for each model

Value

a ggplot2 object

Examples

Run this code

# NOT RUN {
# }
# NOT RUN {
library("breakDown")
library("randomForest")
HR_rf_model <- randomForest(left~., data = breakDown::HR_data, ntree = 100)
explainer_rf  <- explain(HR_rf_model, data = HR_data, y = HR_data$left)
vd_rf <- variable_dropout(explainer_rf, type = "raw")
vd_rf
plot(vd_rf)

HR_glm_model <- glm(left~., data = breakDown::HR_data, family = "binomial")
explainer_glm <- explain(HR_glm_model, data = HR_data, y = HR_data$left)
logit <- function(x) exp(x)/(1+exp(x))
vd_glm <- variable_dropout(explainer_glm, type = "raw",
                        loss_function = function(observed, predicted)
                                   sum((observed - logit(predicted))^2))
vd_glm
plot(vd_glm)

library("xgboost")
model_martix_train <- model.matrix(left~.-1, breakDown::HR_data)
data_train <- xgb.DMatrix(model_martix_train, label = breakDown::HR_data$left)
param <- list(max_depth = 2, eta = 1, silent = 1, nthread = 2,
              objective = "binary:logistic", eval_metric = "auc")
HR_xgb_model <- xgb.train(param, data_train, nrounds = 50)
explainer_xgb <- explain(HR_xgb_model, data = model_martix_train,
                                    y = HR_data$left, label = "xgboost")
vd_xgb <- variable_dropout(explainer_xgb, type = "raw")
vd_xgb
plot(vd_xgb)

plot(vd_rf, vd_glm, vd_xgb)

# NOTE:
# if you like to have all importances hooked to 0, you can do this as well
vd_rf <- variable_dropout(explainer_rf, type = "difference")
vd_glm <- variable_dropout(explainer_glm, type = "difference",
                        loss_function = function(observed, predicted)
                                   sum((observed - logit(predicted))^2))
vd_xgb <- variable_dropout(explainer_xgb, type = "difference")
plot(vd_rf, vd_glm, vd_xgb)
# }
# NOT RUN {
# }

Run the code above in your browser using DataCamp Workspace