R语言：如何修改月度最大值代码以计算前N个最大值的平均值？

Hey there! Let's adjust your existing R code to compute the average of the top 2 or 3 monthly maximum values instead of just the single monthly max. Below are two easy-to-follow approaches, including a base R method that builds on your original code, plus a tidyverse/dplyr approach that's great for readability.

Base R Approach (Building on Your Original Code) #

First, we'll create a custom function that takes a vector, filters out missing values, sorts it, picks the top N largest values, and returns their average. This function works directly with aggregate():

# Define a custom function to calculate average of top N values
top_n_mean <- function(x, n = 2) {
  # Remove NA values to avoid calculation errors
  x_clean <- x[!is.na(x)]
  # Return NA if there aren't enough valid values for the top N
  if (length(x_clean) < n) {
    return(NA)
  }
  # Sort values ascending, take last N (largest) values, then compute mean
  mean(tail(sort(x_clean), n))
}

# First, ensure your date column is formatted correctly (same as your original code)
DF$date <- as.Date(DF$date, format = "%Y-%m-%d")

# Calculate average of top 2 monthly values
Output_top2 <- aggregate(
  DF[,-1], 
  by = list(Month = format(DF$date, "%y-%m")), 
  FUN = top_n_mean, 
  n = 2  # Specify number of top values here
)

# Calculate average of top 3 monthly values (just change the n parameter)
Output_top3 <- aggregate(
  DF[,-1], 
  by = list(Month = format(DF$date, "%y-%m")), 
  FUN = top_n_mean, 
  n = 3
)

Tidyverse/dplyr Approach (More Intuitive for Beginners) #

If you're open to using the dplyr package (part of the tidyverse), this method is often easier to read and modify. We'll use group_by() to group by month, then summarise(across()) to apply our calculation to all non-date columns:

# Load required packages
library(dplyr)

# Clean and prepare your data
DF <- DF %>%
  mutate(
    date = as.Date(date, format = "%Y-%m-%d"),
    Month = format(date, "%y-%m")  # Create monthly grouping column
  )

# Average of top 2 monthly values
Output_top2_dplyr <- DF %>%
  group_by(Month) %>%
  summarise(
    across(
      -date,  # Apply to all columns except date
      ~mean(tail(sort(.[!is.na(.)]), 2), na.rm = FALSE)
    )
  )

# Average of top 3 monthly values (replace 2 with 3)
Output_top3_dplyr <- DF %>%
  group_by(Month) %>%
  summarise(
    across(
      -date,
      ~mean(tail(sort(.[!is.na(.)]), 3), na.rm = FALSE)
    )
  )

Key Notes: #

• Handling Missing Values: Both methods include checks for NA values to prevent unexpected errors. If a month has fewer than N valid data points, it will return NA for that month's average (you can adjust this behavior if needed, e.g., return the mean of available values).
• Sorting Logic: sort(x) orders values in ascending order, so tail(sort(x), n) grabs the last N values, which are the largest N values in the vector.

内容的提问来源于stack exchange，提问作者Samima