R语言决策树calibration plot代码

以下是R语言决策树calibration plot的代码：

library(rpart.plot)
library(ggplot2)

# 构建决策树模型
model <- rpart(Species ~ ., data = iris)

# 预测结果
pred <- predict(model, type = "prob")

# 将预测结果与实际结果合并
result <- data.frame(pred, iris$Species)

# 计算每个预测概率区间的实际结果比例
calib <- result %>%
  group_by(Species) %>%
  mutate(bin = cut(pred[, Species], breaks = seq(0, 1, 0.1), include.lowest = TRUE)) %>%
  group_by(bin, .add = TRUE) %>%
  summarise(n = n()) %>%
  mutate(prop = n / sum(n))

# 绘制calibration plot
ggplot(calib, aes(x = as.numeric(bin), y = prop, fill = Species)) +
  geom_bar(stat = "identity", position = "dodge") +
  scale_fill_manual(values = c("#00AFBB", "#E7B800", "#FC4E07")) +
  labs(x = "Predicted probability", y = "Actual proportion", fill = "Species") +
  theme_minimal()

希望对你有帮助！