I have numeric vectors with different lengths, ranging from 300 to 500. I would like to 'normalize' them to a length of 100, i.e. for a vector of length 300 I take the mean of 3 values, for a vector of length 500 the mean of 5 values and so on.
How can I bin numeric vectors and calculate the mean without reordering? I have not been successful with cut so far.
# numeric vectors of different lengths
v1 = rnorm(300)
v2 = rnorm(500)
# goal: numeric vectors of same length
v1.binned = c(mean(v1[1],v1[2],v1[3]), ...)
v2.binned = c(mean(v2[1],v2[2],v2[3], v2[4], v2[5]), ...)
You can convert the vectors to a matrix and use colMeans:
colMeans(matrix(v1,100))
[1] -0.09583398 0.01330998 0.11107002
colMeans(matrix(v2,100))
[1] -0.02396420 0.08638535 -0.03953273 0.09861287 0.01112838
Though beware of recycling if the cut size is not an exact multiple of the vector size. In which case, a split-sapply strategy will do the job:
sapply(split(v1,(seq_along(v1)-1)%/%200),mean)
0 1
-0.041262 0.111070