scale1 function

A variation over the base R scale function that avoids the "divide by 0 = NA" problem.

scale1: A variation over the base R scale function. The function scale1: centers (if needed) and scales a vector to norm 1; if the vector contains values all equal to a constant, scale1 sets all values to 0 (in lieu of NA as scale does). Usefull when pre-processing tables for PCA-like type of analyses.

scale1(x, scale = TRUE, center = TRUE)

Arguments

• x: a vector to be scaled

• scale: (default = TRUE), when TRUE

  scale the vector to norm 1, otherwise do nothing.

• center: (default = TRUE), when TRUE

  center the vectors (i.e., substract the mean from all numbers), otherwise do nothing.

Returns

a centered (if required) and norm-1 (if required) normalized vector.

Examples

toto   <- runif(10)     # 10 random numbers between 0 and 1
 tutu   <- scale1(toto)  # toto centered and normalized
 toto0  <- rep(1,10)     # 10 numbers all equal to 1
 tutu0  <- scale1(toto0) # scaled to 0 # Compare with
 tutuNA <- scale(toto0)  # all numbers set to NA

Author(s)

Hervé Abdi