如何移除满足特定百分比元素小于指定数值的矩阵行

Hey there! Let's adjust your existing function to filter out rows where more than X% of the numeric columns are below a specified value (like 75% below 10). Here's a clear, modified version with explanations:

Modified Function #

First, let's update the function to include a percentage threshold parameter (pct) and calculate the proportion of values below your target number (n) per row:

filter_rows <- function(m, n, pct) {
  # Calculate the percentage of numeric columns below 'n' for each row
  below_threshold_pct <- apply(m[, -1], 1, function(row) mean(row < n))
  
  # Keep rows where the percentage of values below 'n' is <= pct
  m[below_threshold_pct <= pct, ]
}

Key Changes Explained #

• We replaced the all(...) check with mean(row < n): Since logical values (TRUE/FALSE) are treated as 1/0 in arithmetic operations, taking the mean gives us the proportion of values below n in that row.
• Added the pct parameter: This lets you specify the threshold (e.g., 0.75 for 75%) to filter against.
• m[, -1] targets only the numeric columns (skipping the first character column) — works for both data frames and matrices.

Example Usage with Your Sample Data #

Let's test this with your example matrix:

# Create your sample matrix/data frame
sample_mat <- data.frame(
  V1 = c("E.5", "E.2", "E.6", "E.1", "E.3"),
  V2 = c(13, 7, 7, 30, 51),
  V3 = c(16, 3, 7, 18, 54),
  V4 = c(19, 3, 6, 28, 52),
  V5 = c(7, 5, 6, 16, 44),
  stringsAsFactors = FALSE
)

# Filter out rows where >75% of numeric values are below 10
filtered_result <- filter_rows(sample_mat, n = 10, pct = 0.75)
print(filtered_result)

Output #

V1 V2 V3 V4 V5
1 E.5 13 16 19  7
4 E.1 30 18 28 16
5 E.3 51 54 52 44

As you can see:

• Rows E.2 and E.6 are filtered out: 100% of their numeric values are below 10 (which exceeds the 75% threshold).
• Rows E.5, E.1, and E.3 are kept: their proportion of values below 10 is 25%, 0%, and 0% respectively — all ≤75%.

内容的提问来源于stack exchange，提问作者Fate