mean_n: Mean of Numeric Values

In dvmisc: Faster Computation of Common Statistics and Miscellaneous Functions

Description

Defined simply as sum(x) / length(x), this function seems to always run faster than mean for numeric, non-integer vectors/matrices. For integer objects, mean_i should run even faster.

Usage

mean_n(x)

Arguments

x	Numeric vector or matrix.

Value

Numeric value.

Examples

# In general, mean_n is much faster than mean.
x <- rnorm(10000)
mean(x) == mean_n(x)
benchmark(mean(x), mean_n(x), replications = 1000)

# For very large integer objects, mean may be faster than mean_n. But then 
# mean_i should be even faster.
x <- rpois(100000, lambda = 5)
mean(x) == mean_n(x)
mean(x) == mean_i(x)
benchmark(mean(x), mean_n(x), mean_i(x), replications = 1000)

dvmisc documentation built on May 2, 2019, 5:51 p.m.