R 中矩阵列表的总和

Question

我试图将矩阵列表放在一个列表中，然后在每个列表中求和。以下是代码的简单示例：

假设我有 4 个矩阵：

x1 <- matrix(1:9, nrow = 3)
x2 <- matrix(2:10, nrow = 3)
x3 <- matrix(3:11, nrow = 3)
x4 <- matrix(4:12, nrow = 3)

我想以这样的方式将它们放入

list()

中：

[[1]]
[[1]][[1]]
     [,1] [,2] [,3]
[1,]    1    4    7
[2,]    2    5    8
[3,]    3    6    9

[[1]][[2]]
     [,1] [,2] [,3]
[1,]    2    5    8
[2,]    3    6    9
[3,]    4    7   10


[[2]]
     [,1] [,2] [,3]
[1,]    3    6    9
[2,]    4    7   10
[3,]    5    8   11

[[3]]
     [,1] [,2] [,3]
[1,]    4    7   10
[2,]    5    8   11
[3,]    6    9   12

如何对

list()

内的每个元素进行求和？ 例如，我想要的输出如下：

[[1]]
     [,1] [,2] [,3]
[1,]    3    9   15
[2,]    5   11   17
[3,]    7   13   19

[[2]]
     [,1] [,2] [,3]
[1,]    3    6    9
[2,]    4    7   10
[3,]    5    8   11

[[3]]
     [,1] [,2] [,3]
[1,]    4    7   10
[2,]    5    8   11
[3,]    6    9   12

我尝试过使用

list(Reduce(`+`, x))

但是它不起作用。

Answer 1

既然你想保留顶级列表，请使用

lapply

:

lapply(x, function(l) if(is.list(l)) Reduce(`+`, l) else l)

#[[1]]
#     [,1] [,2] [,3]
#[1,]    3    9   15
#[2,]    5   11   17
#[3,]    7   13   19

#[[2]]
#     [,1] [,2] [,3]
#[1,]    3    6    9
#[2,]    4    7   10
#[3,]    5    8   11

#[[3]]
#     [,1] [,2] [,3]
#[1,]    4    7   10
#[2,]    5    8   11
#[3,]    6    9   12

Answer 2

对应的

purrr

版本的@RonakShah的答案与

map_if()

：

library(purrr)

map_if(x, is.list, reduce, `+`)

# [[1]]
#      [,1] [,2] [,3]
# [1,]    3    9   15
# [2,]    5   11   17
# [3,]    7   13   19
# 
# [[2]]
#      [,1] [,2] [,3]
# [1,]    3    6    9
# [2,]    4    7   10
# [3,]    5    8   11
# 
# [[3]]
#      [,1] [,2] [,3]
# [1,]    4    7   10
# [2,]    5    8   11
# [3,]    6    9   12

Answer 3

对我来说，使用

array

更优雅。这是一些示例代码。该示例计算平均值而不是总和。

R <- list(matrices, all of the same dimension)
# Combine all these into an array with an extra dimension
# corresponding to the different matrices
dn <- dimnames(R[[1]])   # fetch row and column names
rn <- dn[[1]]; cn <- dn[[2]]
A <- array(unlist(R), dim=c(length(rn), length(cn), length(R)),
           dimnames=list(rn, cn, NULL))
# Compute mean (separately for each metrix cell) 
# over matrices, ignoring NAs
Rmean <- apply(R, 1:2, mean, na.rm=TRUE)