从R中for循环的同一迭代中提取不同的统计信息答案

【问题标题】：Extracting different statistics from the same iteration of a for loop in R从R中for循环的同一迭代中提取不同的统计信息
【发布时间】：2021-01-29 02:34:55
【问题描述】：

我想在两个变量之间创建一个数据框，这些变量链接到每次迭代的相同分布。

sim_df <- rep(NA, 1000)
for (i in 1:1000) { 

#randomly generated distribution
dist = sample(c(0:200), 10, replace = TRUE)
dist = as.data.frame(dist)

#calculated statistics
dist_mean = mean(a[,1])
dist_sd   = sd(a[,1])

#build data frame
df = data.frame(dist_mean, dist_sd)

#return
sim_df = df[i]
}

我不确定如何定义每一列以便提取它们。

【问题讨论】：

标签： r dataframe loops for-loop

【解决方案1】：

创建一个函数，该函数在一次迭代中返回均值和 sd。

return_mean_sd <- function() {
  dist = sample(c(0:200), 10, replace = TRUE)
  c(mean = mean(dist), sd = sd(dist))
}

您可以调用此函数n 次来创建数据框。

set.seed(123)
sim_df <- as.data.frame(t(replicate(100, return_mean_sd())))
head(sim_df)

#   mean   sd
#1 105.0 70.3
#2 119.7 44.3
#3  91.6 56.1
#4  91.5 58.7
#5  99.1 55.2
#6  90.0 52.1

【讨论】：

【解决方案2】：

我认为这就是你想要做的：

set.seed(2021) # for reproducibility, do not use in production
sim_df <- replicate(1000, {
  dist = sample(c(0:200), 10, replace = TRUE)
  data.frame(dist_mean = mean(dist), dist_sd = sd(dist))
}, simplify = FALSE)

sim_df[1:3]
# [[1]]
#   dist_mean  dist_sd
# 1     143.1 40.04012
# [[2]]
#   dist_mean  dist_sd
# 1       119 55.65169
# [[3]]
#   dist_mean  dist_sd
# 1       103 44.23674

从这里开始，如果您希望所有这些都在一帧中，那么

sim_df1 <- do.call(rbind, sim_df)
nrow(sim_df1)
# [1] 1000
head(sim_df1)
#   dist_mean  dist_sd
# 1     143.1 40.04012
# 2     119.0 55.65169
# 3     103.0 44.23674
# 4     103.2 57.00838
# 5     124.3 50.74785
# 6     112.9 56.98041

【讨论】：

【解决方案3】：

更容易生成完整的实现流，将它们转换为矩阵，然后从行中生成统计数据：

dist <- sample(c(0:200), 10 * 1000, replace = TRUE)
dist <- matrix(dist, ncol = 10)
means <- rowMeans(dist)
sds <- apply(dist, 1, sd)
df <- data.frame(means, sds)
head(df)
  means      sds
1  60.7 52.45750
2 124.3 65.82646
3 101.3 73.34856
4 116.4 39.37907
5  97.2 65.79227
6  94.7 58.27530

【讨论】：