Generic Function for Breaking Down of Model Predictions
The broken function is a generic function for decomposition of model predictions.
For linear models please use broken.lm, for generic linear models please use broken.glm.
For all other models please use the model agnostic version broken.default.
Please note, that some of these functions have additional parameters.
broken(model, new_observation, ...)
model |
a model |
new_observation |
a new observation with columns that corresponds to variables used in the model |
... |
other parameters |
an object of the broken class
## Not run:
library("breakDown")
library("randomForest")
library("ggplot2")
set.seed(1313)
model <- randomForest(factor(left)~., data = HR_data, family = "binomial", maxnodes = 5)
predict.function <- function(model, new_observation)
predict(model, new_observation, type="prob")[,2]
predict.function(model, HR_data[11,-7])
explain_1 <- broken(model, HR_data[11,-7], data = HR_data[,-7],
predict.function = predict.function, direction = "down")
explain_1
plot(explain_1) + ggtitle("breakDown plot (direction=down) for randomForest model")
explain_2 <- broken(model, HR_data[11,-7], data = HR_data[,-7],
predict.function = predict.function, direction = "down", keep_distributions = TRUE)
plot(explain_2, plot_distributions = TRUE) +
ggtitle("breakDown distributions (direction=down) for randomForest model")
explain_3 <- broken(model, HR_data[11,-7], data = HR_data[,-7],
predict.function = predict.function, direction = "up", keep_distributions = TRUE)
plot(explain_3, plot_distributions = TRUE) +
ggtitle("breakDown distributions (direction=up) for randomForest model")
## End(Not run)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.