R语言 对LightGBM回归预测模型计算变量重要性，绘制横向对称柱状图，纵轴左侧为Gain，纵轴右侧为Cover，横轴为数值大小，纵轴排列Feature，背景为白色，全部字体设置为"times new roman"

在R语言中，使用LightGBM进行回归预测并计算变量的重要性通常涉及到`lightgbm`包以及数据可视化库，如`ggplot2`或`grDevices`。要创建一个包含Gain和Cover信息的横向对称柱状图，你可以按照以下步骤操作：

1. 首先，加载必要的库：

library(lightgbm)
library(ggplot2)

2. 训练LightGBM模型并获取变量重要性数据：

lgb_train <- lgb.Dataset(data = your_data, label = your_labels) # 替换为实际的数据和标签
params <- list(objective = "regression", # 如果是回归任务
              num_leaves = 31, # 可调整的参数
              learning_rate = 0.05,
              min_child_weight = 10,
              importance_type = "gain") # 获取Gain值

model <- lgb.train(params, lgb_train)
importance <- varImp(model)

3. 创建数据框来存储重要性数据：

imp_df <- data.frame(FEATURE = rownames(importance), Gain = importance$gain, Cover = importance$cover)

4. 使用`ggplot2`绘制图形：

ggplot(imp_df, aes(x = NULL, y = reorder(FEATURE, -Gain), fill = c(Gain, Cover))) +
  geom_bar(stat = "identity", width = unit(1, "npc"), position = position_dodge(width = 0.9)) +
  scale_fill_manual(values = c("white", "#E69F00")) + # Gain为黄色，Cover为橙色
  geom_text(aes(label = round(Gain, 2), color = "black"),
            hjust = 1, vjust = 1, size = 3, family = "Times New Roman") +
  geom_text(aes(label = round(Cover, 2), color = "black"),
            hjust = -1, vjust = 1, size = 3, family = "Times New Roman") +
  theme_minimal(base_size = 12) +
  coord_flip() +
  labs(title = "Variable Importance (Gain and Cover)", x = "", y = "") +
  guides(fill = guide_legend(reverse = TRUE)) +
  theme(plot.background = element_rect(fill = "white", color = NA),
        axis.title.y = element_blank(),
        axis.text.y = element_text(family = "Times New Roman"),
        axis.ticks.y = element_blank()) # 设置字体等样式

记得将`your_data`和`your_labels`替换为你实际的数据和目标变量。