# Required packages
# install.packages(c("randomForest", "readxl", "writexl", "ggplot2", "dplyr"))

library(randomForest)
library(readxl)
library(writexl)
library(ggplot2)
library(dplyr)

# Read data
features_df <- read_excel("C:/Users/12145/X_train.xlsx")
labels_df   <- read_excel("C:/Users/12145/y_train.xlsx")

# Prepare data
X_train <- as.data.frame(features_df)
y_train <- as.factor(unlist(labels_df))

# Train random forest model
model <- randomForest(x = X_train, y = y_train, importance = TRUE)

# Get feature importance
feature_importances <- importance(model, type = 1)
importance_df <- data.frame(
  Feature = rownames(feature_importances),
  Importance = feature_importances[, 1]
)

# Sort by importance
importance_df <- importance_df %>%
  arrange(Importance)

# Plot feature importance
ggplot(importance_df, aes(x = Importance, y = reorder(Feature, Importance))) +
  geom_bar(stat = "identity", fill = "#4C9F70") +
  theme_minimal() +
  theme(
    axis.text.y = element_text(size = 6),
    axis.text.x = element_text(size = 8),
    plot.title = element_text(size = 14, face = "bold")
  ) +
  labs(
    title = "Feature Importance from Random Forest (Sorted)",
    x = "Feature Importance",
    y = "Features"
  )

# Save to Excel
output_path <- "C:/Users/12145/feature_importances.xlsx"
write_xlsx(importance_df, output_path)

cat("Feature importance file saved to:", output_path, "\n")