Calculating means of subsequences using loop-CodePudding

I've used rnorm to simulate n=100 draws from a normal. What I want to do now is to calculate the mean of "subsequences" of the data, i.e. mean of elements 1:10, 1:20, 1:30, ..., 1:100.

How can I do that using loop which just saves the calculated means instead of first creating subsets and consequently calculating means?

CodePudding user response：

Create a splitting vector f using a cumsum trick and tapply or aggregate with function mean.

set.seed(2022)
x <- rnorm(100)

k <- 10L
f <- c(1, rep(0, k - 1L))
f <- rep(f, length.out = length(x))
f <- cumsum(f)

tapply(x, f, mean)
#>            1            2            3            4            5            6 
#> -0.563706565  0.007282962  0.208630598  0.063939372 -0.360526835  0.622263561 
#>            7            8            9           10 
#> -0.096927090  0.753811231  0.462543860  0.290149022

aggregate(x ~ f, FUN = mean)
#>     f            x
#> 1   1 -0.563706565
#> 2   2  0.007282962
#> 3   3  0.208630598
#> 4   4  0.063939372
#> 5   5 -0.360526835
#> 6   6  0.622263561
#> 7   7 -0.096927090
#> 8   8  0.753811231
#> 9   9  0.462543860
#> 10 10  0.290149022

^{Created on 2022-04-08 by the reprex package (v2.0.1)}

Edit

I misread the problem, like it is said in a OP comment,

Thanks! The problem is that this calculates means from 1:10, 11:20, 21:30 but I would actually need 1:10, 1:20, 1:30 until all 100 elements

g <- seq(10L, length(x), by = 10L)
sapply(g, \(k) mean(x[seq_len(k)]))
#>  [1] -0.563706565 -0.278211802 -0.115931002 -0.070963408 -0.128876094
#>  [6] -0.003686151 -0.017006285  0.079345904  0.121923455  0.138746012

^{Created on 2022-04-08 by the reprex package (v2.0.1)}

CodePudding user response：

set.seed(2022)
x <- rnorm(100)
mn <- colMeans(matrix(x, 10))
cumsum(mn)/seq(mn)

[1] -0.563706565 -0.278211802 -0.115931002 -0.070963408 -0.128876094
[6] -0.003686151 -0.017006285  0.079345904  0.121923455  0.138746012