library("randomForest")
# First we fit a machine learning model on the Boston housing data
data("Boston", package = "MASS")
X <- Boston[-which(names(Boston) == "medv")]
rf <- randomForest(medv ~ ., data = Boston, ntree = 50)
mod <- Predictor$new(rf, data = X)
# Explain the first instance of the dataset with the LocalModel method:
x.interest <- X[1, ]
lemon <- LocalModel$new(mod, x.interest = x.interest, k = 2)
predict(lemon, newdata = x.interest)
Run the code above in your browser using DataLab