修剪 R 中的极值

Question

I'm working on a large data and I'm trying to trim the extreme values like the ones < 0 and > 8 how can I do that in R?

Atrim = c(0,8)

Answer 1

You were not clear how to handle the outliers. Should they be elimnated or changed to NAs? Here is a simple example that deletes the extreme values:

set.seed(42) # for reproducibility
x <- rnorm(50, 4, 2.5)   # Generate 50 random values
y <- x[x >= 0 & x <= 8]  # Remove outliers
length(y)
# [1] 41  # Nine values were removed.

If the data are in a matrix and you want to replace the outliers with NA, try this:

set.seed(42) # for reproducibility
x <- matrix(rnorm(49, 4, 2.5), 7, 7)   # Generate 7 x7 matrix of random values
print(x, digits=3)  # Print with 3 significant digits
#      [,1]  [,2]  [,3]   [,4] [,5]   [,6]  [,7]
# [1,] 7.43 3.763  3.67 -0.453 5.15 -0.293 5.895
# [2,] 2.59 9.046  5.59  3.570 2.40  2.039 2.183
# [3,] 4.91 3.843  3.29  7.037 5.14  1.873 0.579
# [4,] 5.58 7.262 -2.64  8.738 5.76 -2.036 5.082
# [5,] 5.01 9.717 -2.10  2.924 6.59  4.090 1.972
# [6,] 3.73 0.528  7.30  3.357 2.48  4.515 7.610
# [7,] 7.78 3.303  3.23 -0.408 5.26  3.097 2.921
idx <- which(x < 0 | x > 8, arr.ind=TRUE)   # Identify outliers
x[idx] <- NA
print(x, digits=3)
#      [,1]  [,2] [,3] [,4] [,5] [,6]  [,7]
# [1,] 7.43 3.763 3.67   NA 5.15   NA 5.895
# [2,] 2.59    NA 5.59 3.57 2.40 2.04 2.183
# [3,] 4.91 3.843 3.29 7.04 5.14 1.87 0.579
# [4,] 5.58 7.262   NA   NA 5.76   NA 5.082
# [5,] 5.01    NA   NA 2.92 6.59 4.09 1.972
# [6,] 3.73 0.528 7.30 3.36 2.48 4.51 7.610
# [7,] 7.78 3.303 3.23   NA 5.26 3.10 2.921