Nested loop in R with two level of variation - r

Question

Hello dear stack overflow community,
Here is the context of my problem : I have a dataframe with each column corresponding to one bat species and and each row corresponds to the acoustic activity measured for one night (for each night of recording not all the species as been sampled).
eg :
> Dataset
Bba Ese Hsa Mda Mda.Mca Mema Mpu
1 3 NA NA NA 33 NA NA
2 NA NA NA NA 1 NA NA
3 2 4 1 NA 19 1 NA
4 NA NA NA NA 25 NA NA
5 NA NA NA NA 3 NA NA
6 1 1 NA NA 53 NA NA
7 1 NA 9 NA NA 1 NA
8 NA NA 10 NA NA NA NA
9 NA NA NA NA NA NA NA
10 1 1 NA NA NA NA NA
11 6 NA NA NA NA NA NA
12 12 NA 1 NA NA 1 NA
13 3 NA 2 NA NA 1 NA
14 1 NA NA NA NA NA NA
15 NA NA NA NA NA NA NA
16 1 NA NA NA NA NA NA
17 2 NA NA NA NA 2 NA
18 1 1 NA NA NA NA 1
19 NA NA NA NA NA NA NA
20 1 1 NA NA NA NA NA
21 2 NA 1 NA NA NA NA
22 1 NA NA NA NA 4 NA
23 1 NA 1 NA NA 1 NA
24 NA NA NA NA NA 2 NA
25 1 NA NA NA NA NA NA
26 1 NA NA NA NA 1 NA
27 1 NA NA NA NA NA NA
28 5 NA NA NA NA NA NA
29 NA NA NA NA NA NA NA
.....
To study vocal activity I am checking the quantile of bat vocal activity per species
apply(Dataset[,9:15],2,quantile, na.rm=TRUE, type=7, c(0.02,0.25,0.5,0.75,0.98))
Bba Ese Hsa Mda Mda.Mca Mema Mpu
2% 1.00 1.00 1.00 1.00 1.00 1.00 1
25% 1.00 1.00 2.00 2.00 2.00 1.00 1
50% 3.00 4.00 6.00 4.00 3.00 2.00 2
75% 9.75 12.00 18.00 12.00 20.00 4.00 6
98% 53.86 69.88 166.12 313.32 159.04 27.28 44
To test the impact of sampling (number of night) on my quantile estimate, I want to do a boostrap. More specifically, I want to calculate the mean of the bat activity if I take only 3 night per species using 1000 random sample with replacement. And i want to do it If I take from 3 to 70 nights.
This is what I have so far (for one species):
Bbana<-as.data.frame(Bbana)
L= length(Bbana[,1])
B= 1000
m<-list()
for (j in 3:70) {
for (i in 1 : B) {
idx<-sample(1:L, j, replace=TRUE)
data_idx<-Bbana[idx, ]
m[i]<-mean(data_idx)
}}
Somehow it didn't give my what I am expected : 67 list with 1000 means of bat activity.
Could anyone help me ?
(I don't know if it's clear enough...)
Thanks in advance

Answer 1

if you want to stick to loops and lists:
for (j in 3:70) {
mat = matrix(NA, nrow = B, ncol = ncol(idx))
for (i in 1 : B) {
idx<-sample(1:L, j, replace=TRUE)
data_idx<-Bbana[idx, ]
mat[i,] = colMeans(data_idx, na.rm = TRUE)
}
m[[j]] = mat
}
Otherwise, this option should work (and should be more efficient / convenient to use):
sample.fun = function(nb.nights, dataset){
# select randomly nb.nights rows to sample
selected.rows = sample(1:nrow(dataset), nb.nights, replace = FALSE)
# return a vector with their means
return(colMeans(dataset[select.rows,], na.rm = TRUE))
}
sapply(3:67, function(nights) replicate(1000, sample.fun(nights, dataset), simplify = 'array'), simplify = FALSE)
This will return you a list of 67 elements that each contains a dataframe of 1000 rows (1000 means per species)