个人中心
gpt4 book ai didi

r - 根据条件创建累积每周平均值的 Date.table 解决方案

转载 作者：行者123 更新时间：2023-12-04 02:31:25 24 4
gpt4 key购买 nike

我每天测量 x 数据。每个唯一 ID (a, b) 都有多行数据。对于每个唯一 ID，我想计算第 1-2 周的每周平均值和第 3 周的累计每周平均值 - (结束日期)属于 Start.date 和 End.date 的 x 数据。每个每周平均值将是一个新列。

以下代码通过为每个唯一 ID 创建一个周变量来计算每周平均值:

dcast(dat[,
                  week := floor(difftime(Date, Start.date, units = "weeks")) + 1,
                  by = .(ID)][,
                              .(weekly_mean = mean(x)),
                              by = .(ID, week)],
      ID ~ paste("week", week, sep = "_"),
      value.var = "weekly_mean")

我可以根据现有的每周平均值创建这些累积的每周平均值吗？如果是这样，一旦我有了累积平均值列，我最终是否可以删除第 3 周(结束日期)的每周平均值？

我希望生成的数据集看起来如何:

Date x Start.Date End.Date ID. Week_1 Week_2 Week_3_cum Week_4_cum ...

一些数据:


dat <- structure(list(Date = structure(c(1104969600, 1105056000, 
1105142400, 1105228800, 1105315200, 1105401600, 1105488000, 1105574400, 
1105660800, 1105747200, 1105833600, 1105920000, 1106006400, 1106092800, 
1106179200, 1106265600, 1106352000, 1106438400, 1106524800, 1106611200, 
1106697600, 1106784000, 1106870400, 1106956800, 1107043200, 1107129600, 
1107216000, 1107302400, 1107388800, 1107475200, 1107561600, 1107648000, 
1107734400, 1107820800, 1107907200, 1107993600, 1108080000, 1108166400, 
1108252800, 1108339200, 1108425600, 1108512000, 1108598400, 1108684800, 
1108771200, 1108857600, 1108944000, 1109030400, 1109116800, 1109203200, 
1109289600, 1109376000, 1109462400, 1109548800, 1104969600, 1105056000, 
1105142400, 1105228800, 1105315200, 1105401600, 1105488000, 1105574400, 
1105660800, 1105747200, 1105833600, 1105920000, 1106006400, 1106092800, 
1106179200, 1106265600, 1106352000, 1106438400, 1106524800, 1106611200, 
1106697600, 1106784000, 1106870400, 1106956800, 1107043200, 1107129600, 
1107216000, 1107302400, 1107388800, 1107475200, 1107561600, 1107648000, 
1107734400, 1107820800, 1107907200, 1107993600, 1108080000, 1108166400, 
1108252800, 1108339200, 1108425600, 1108512000, 1108598400, 1108684800, 
1108771200, 1108857600, 1108944000, 1109030400, 1109116800, 1109203200, 
1109289600, 1109376000, 1109462400, 1109548800), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), x = c(9.5, 9.4, 9.7, 10.11, 11.11, 
12.11, 13.11, 14.11, 15.11, 10.11, 11.11, 12.11, 13.11, 14.11, 
15.11, 9.5, 9.4, 9.7, 10.11, 11.11, 12.11, 13.11, 14.11, 15.11, 
10.11, 11.11, 12.11, 13.11, 14.11, 15.11, 9.5, 9.4, 9.7, 10.11, 
11.11, 12.11, 13.11, 14.11, 15.11, 10.11, 11.11, 12.11, 13.11, 
14.11, 15.11, 9.5, 9.4, 9.7, 10.11, 11.11, 12.11, 13.11, 14.11, 
15.11, 10.11, 11.11, 12.11, 13.11, 14.11, 15.11, 9.5, 9.4, 9.7, 
10.11, 11.11, 12.11, 13.11, 14.11, 15.11, 10.11, 11.11, 12.11, 
13.11, 14.11, 15.11, 9.5, 9.4, 9.7, 10.11, 11.11, 12.11, 13.11, 
14.11, 15.11, 10.11, 11.11, 12.11, 13.11, 14.11, 15.11, 9.5, 
9.4, 9.7, 10.11, 11.11, 12.11, 13.11, 14.11, 15.11, 16.11, 17.11, 
18.11, 19.11, 20.11, 21.11, 22.11, 23.11, 24.11), Start.Date = structure(c(1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 
1104969600, 1104969600, 1104969600, 1104969600, 1104969600, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 1104883200, 
1104883200, 1104883200, 1104883200, 1104883200, 1104883200), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), End.Date = structure(c(1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 
1109635200, 1109635200, 1109635200, 1109635200, 1109635200, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 1109548800, 
1109548800, 1109548800, 1109548800, 1109548800, 1109548800), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), ID = c("a", "a", "a", "a", "a", "a", 
"a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", 
"a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", 
"a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", "a", 
"a", "a", "a", "a", "a", "a", "a", "a", "a", "b", "b", "b", "b", 
"b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", 
"b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", 
"b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b", 
"b", "b", "b", "b", "b", "b", "b", "b", "b", "b", "b")), row.names = c(NA, 
-108L), class = c("tbl_df", "tbl", "data.frame"))

setDT(dat)

最佳答案

我相信这是一种方法。

请注意，虽然您提到了累积均值，但 cummean() 在 R 中意味着某些东西。cummean() 返回与输入长度相同的向量，因此不会聚合成一个值。所以步骤是:

  1. 聚合以找到每个组的计数和sum(O3)
  2. 在前两周，简单地求出每周的平均值。
  3. 对于剩余的几周，使用 cumsum 计算运行平均值。
  4. dcast 到您的预期结果中。
library(data.table)

dat[,week:= floor(difftime(AQ.Date, Start.Date, uni = "weeks")) + 1, by = ID]

ans = dat[,
    .(.N, weekly_sum = sum(O3)),
    by = .(ID, week)
    ][, 
      {
        lgl_ind = week < 3
        first_weeks = which(lgl_ind)
        last_weeks = which(!lgl_ind)
        
        out = numeric(length(week))
        out[first_weeks] = weekly_sum[first_weeks] / N[first_weeks]
        out[last_weeks] = cumsum(weekly_sum[last_weeks]) / cumsum(N[last_weeks])
        
        calc_type = c(rep('mean', length(first_weeks)),
                      rep('cummean', length(last_weeks)))
        
        list(week, weekly_mean = out, calc_type)
      },
      by = ID]  

ans
#>         ID       week weekly_mean calc_type
#>     <char> <difftime>       <num>    <char>
#>  1:      a    1 weeks    10.72000      mean
#>  2:      a    2 weeks    12.82429      mean
#>  3:      a    3 weeks    11.00571   cummean
#>  4:      a    4 weeks    11.84357   cummean
#>  5:      a    5 weeks    11.65952   cummean
#>  6:      a    6 weeks    11.87929   cummean
#>  7:      a    7 weeks    11.81886   cummean
#>  8:      a    8 weeks    11.98025   cummean
#>  9:      b    1 weeks    12.61000      mean
#> 10:      b    2 weeks    10.72000      mean
#> 11:      b    3 weeks    12.82429   cummean
#> 12:      b    4 weeks    11.91500   cummean
#> 13:      b    5 weeks    12.17048   cummean
#> 14:      b    6 weeks    11.95071   cummean
#> 15:      b    7 weeks    12.58257   cummean
#> 16:      b    8 weeks    13.90366   cummean

dcast(ans,
      ID ~ paste("week", week,calc_type,  sep = "_"),
      value.var = 'weekly_mean')
#>        ID week_1_mean week_2_mean week_3_cummean week_4_cummean week_5_cummean
#>    <char>       <num>       <num>          <num>          <num>          <num>
#> 1:      a       10.72    12.82429       11.00571       11.84357       11.65952
#> 2:      b       12.61    10.72000       12.82429       11.91500       12.17048
#>    week_6_cummean week_7_cummean week_8_cummean
#>             <num>          <num>          <num>
#> 1:       11.87929       11.81886       11.98025
#> 2:       11.95071       12.58257       13.90366

关于r - 根据条件创建累积每周平均值的 Date.table 解决方案，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/63975552/

24 4 0
文章推荐： algorithm - 找到第 k 个好数
文章推荐： youtube-api - 为什么 Youtube "channels"API 不返回 "contentOwnerDetails"数据？
文章推荐： android-ndk - 如何将采样率从 AV_SAMPLE_FMT_FLTP 转换为 AV_SAMPLE_FMT_S16？
文章推荐： python - 模块未找到错误 : No module named 'SessionState
行者123
个人简介

我是一名优秀的程序员,十分优秀！

滴滴打车优惠券免费领取
滴滴打车优惠券
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com