如何计算r中矩阵列表的平均值

Question

I have a list like below, which is a list of lists containing matrices(so "ftable" is a list of ten lists and each of the internal lists contains seven matrices). I need to calculate the mean of associated matrices which may also have NA values as well.I have tried several ways but I got errors.

for(i in 1:10){
for(j in 1:7){
ftable[[i]][[j]] <- matrix (x,nrow=8,ncol=8, byrow=TRUE)
}
}

> str(ftable)
List of 10
$ :List of 7
.......
.......

as the result I need to have a list containing seven matrices that each of these matrices are the result of applying mean to ftable[[1]][[i]], ftable[[2]][[i]], ... , ftable[[10]][[i]] and i in 1:7.

I have tried this but I got error:

meanTable <- list()
for (i in 1:7)
meanTable[[i]] <- matrix (0, nrow=8,ncol=8)

> meanTable <- lapply(1:7, function(i) Reduce(mean, list(ftable[[1]][i],ftable[[2]][i],ftable[[3]][i],ftable[[4]][i],ftable[[5]][i],ftable[[6]][i],ftable[[7]][i],ftable[[8]][i],ftable[[9]][i],ftable[[10]][i])))
Error in mean.default(init, x[[i]]) : 
'trim' must be numeric of length one
In addition: Warning message:
In mean.default(init, x[[i]]) :
argument is not numeric or logical: returning NA

one example of the matrices:

> ftable[[1]][[1]]
1     2     3      4      5      6      7      8
1 NA 0.924 0.835 -0.336  0.335 -0.948  0.285  0.749
2 NA    NA 0.772 -0.333  0.333 -0.892  0.127  0.715
3 NA    NA    NA -0.476  0.475 -0.756  0.258  0.749
4 NA    NA    NA     NA -0.999  0.399 -0.150 -0.399
5 NA    NA    NA     NA     NA -0.399  0.151  0.399
6 NA    NA    NA     NA     NA     NA -0.134 -0.715
7 NA    NA    NA     NA     NA     NA     NA  0.144
8 NA    NA    NA     NA     NA     NA     NA     NA

Answer 1

I think this is what you require. The easiest way would be to unlist your outer list and then apply Reduce as follows:

I'll create a variation of the input from user1317221_G

set.seed(45)
mat1 <- matrix(c(sample(10),NA,NA),nrow=2)
matlist1 <- list(mat1,mat1,mat1)
mat2 <- matrix(c(sample(11:20),NA,NA),nrow=2)
matlist2 <- list(mat2,mat2,mat2)
bigmatlist <- list(matlist1,matlist2)

> bigmatlist
# [[1]]
# [[1]][[1]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,]    7    2   10    1    4   NA
# [2,]    3    9    8    5    6   NA
# 
# [[1]][[2]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,]    7    2   10    1    4   NA
# [2,]    3    9    8    5    6   NA
# 
# [[1]][[3]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,]    7    2   10    1    4   NA
# [2,]    3    9    8    5    6   NA
# 
# 
# [[2]]
# [[2]][[1]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,]   14   13   19   17   15   NA
# [2,]   18   20   16   11   12   NA
# 
# [[2]][[2]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,]   14   13   19   17   15   NA
# [2,]   18   20   16   11   12   NA
# 
# [[2]][[3]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,]   14   13   19   17   15   NA
# [2,]   18   20   16   11   12   NA

Now the solution.

# in your case, outer.len = 10 and inner.len = 7
outer.len <- 2
inner.len <- 3
prod.len <- outer.len * inner.len
list.un <- unlist(bigmatlist, recursive = FALSE)
o <- lapply(1:inner.len, function(idx) {
    Reduce('+', list.un[seq(idx, prod.len, by = inner.len)])/outer.len
})

> o
# [[1]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,] 10.5  7.5 14.5    9  9.5   NA
# [2,] 10.5 14.5 12.0    8  9.0   NA
# 
# [[2]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,] 10.5  7.5 14.5    9  9.5   NA
# [2,] 10.5 14.5 12.0    8  9.0   NA
# 
# [[3]]
#      [,1] [,2] [,3] [,4] [,5] [,6]
# [1,] 10.5  7.5 14.5    9  9.5   NA
# [2,] 10.5 14.5 12.0    8  9.0   NA

Answer 2

you mean like this:

 mat1 <- matrix(c(1:10,NA,NA),nrow=2)
 matlist1 <- list(mat1,mat1,mat1)
 bigmatlist <- list(matlist1,matlist1)

 mean(mat1, na.rm=TRUE)
 #[1] 5.5

 sapply(matlist1, function(x) mean(x,na.rm=TRUE))
 #[1] 5.5 5.5 5.5

and a list of lists:

 sapply(bigmatlist,function(x) sapply(x, function(x) mean(x,na.rm=TRUE)) )
#      [,1] [,2]
#[1,]  5.5  5.5
#[2,]  5.5  5.5
#[3,]  5.5  5.5

changing the sapply for lapply where appropriate if you want lists returned.

where [3,][,1] is the mean for matrix three in list 1 ie bigmatlist[[1]][[3]]