计算每个组中相同特定值的百分比。

huangapple
go评论96阅读模式
英文:

Calculate percentage of same specific values per group

问题

  1. 我有以下的数据框 df（以下是`dput`）：
  3.     > df
  4.        group class value
  5.     1      A FALSE     2
  6.     2      A  TRUE     1
  7.     3      A FALSE     1
  8.     4      A FALSE     2
  9.     5      A FALSE     3
  10.     6      B FALSE     4
  11.     7      B FALSE     2
  12.     8      B  TRUE     2
  13.     9      B FALSE     2
  14.     10     B FALSE     6
  15.     11     C  TRUE     5
  16.     12     C FALSE     5
  17.     13     C FALSE     3
  18.     14     C FALSE     5
  19.     15     C FALSE     5
  21. 我想要计算每个组中特定相似值的百分比。每个组中始终有一个`class == TRUE`的值，这意味着我想要计算与`class == TRUE`值相似的值的百分比。如上面的数据框中所示，组A有`class == TRUE`的值为1，并且组A中有两个值为1，因此2/5 = 0.4的值为1。以下是所需的输出：
  23.       group value pct
  24.     1     A     1 0.4
  25.     2     B     2 0.6
  26.     3     C     5 0.8
  28. 所以我想知道是否有人知道如何在R中计算每个组中特定值的百分比？
  30. ***
  32. `dput` 的 df：
  34.     df <- structure(list(group = c("A", "A", "A", "A", "A", "B", "B", "B", 
  35.     "B", "B", "C", "C", "C", "C", "C"), class = c(FALSE, TRUE, FALSE, 
  36.     FALSE, FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, TRUE, FALSE, 
  37.     FALSE, FALSE, FALSE), value = c(2, 1, 1, 2, 3, 4, 2, 2, 2, 6, 
  38.     5, 5, 3, 5, 5)), class = "data.frame", row.names = c(NA, -15L
  39.     ))
英文:

I have the following dataframe df (dput below):

  1. &gt; df
  2.    group class value
  3. 1      A FALSE     2
  4. 2      A  TRUE     1
  5. 3      A FALSE     1
  6. 4      A FALSE     2
  7. 5      A FALSE     3
  8. 6      B FALSE     4
  9. 7      B FALSE     2
  10. 8      B  TRUE     2
  11. 9      B FALSE     2
  12. 10     B FALSE     6
  13. 11     C  TRUE     5
  14. 12     C FALSE     5
  15. 13     C FALSE     3
  16. 14     C FALSE     5
  17. 15     C FALSE     5

I would like to calculate the percentage of specific similar values per group. There is always one value with the class == TRUE, which means I would like to calculate the percentage of values similar to the value with class == TRUE. As you can see in the dataframe above group A has the value 1 with class == TRUE and there are two values with 1 in group A so 2/5 = 0.4 of the values are 1. Here is the desired output:

  1.   group value pct
  2. 1     A     1 0.4
  3. 2     B     2 0.6
  4. 3     C     5 0.8

So I was wondering if anyone knows how to calculate the percentage of specific values per group in R?

dput of df:

  1. df &lt;- structure(list(group = c(&quot;A&quot;, &quot;A&quot;, &quot;A&quot;, &quot;A&quot;, &quot;A&quot;, &quot;B&quot;, &quot;B&quot;, &quot;B&quot;, 
  2. &quot;B&quot;, &quot;B&quot;, &quot;C&quot;, &quot;C&quot;, &quot;C&quot;, &quot;C&quot;, &quot;C&quot;), class = c(FALSE, TRUE, FALSE, 
  3. FALSE, FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, TRUE, FALSE, 
  4. FALSE, FALSE, FALSE), value = c(2, 1, 1, 2, 3, 4, 2, 2, 2, 6, 
  5. 5, 5, 3, 5, 5)), class = &quot;data.frame&quot;, row.names = c(NA, -15L
  6. ))

答案1

得分: 3

你可以这样做：

  1. library(dplyr)
  2. df %>%
  3.   group_by(group) %>%
  4.   summarize(pct = sum(value == value[class == TRUE])/n(),
  5.             value = value[class == TRUE])
  7. # 生成一个 tibble 表格：
  8. #   group   pct value
  9. #   <chr> <dbl> <dbl>
  10. # 1 A       0.4     1
  11. # 2 B       0.6     2
  12. # 3 C       0.8     5
英文:

You could do:

  1. library(dplyr)
  2. df %&gt;%
  3.   group_by(group) %&gt;%
  4.   summarize(pct = sum(value == value[class == TRUE])/n(),
  5.             value = value[class == TRUE])
  7. # A tibble: 3 x 3
  8.   group   pct value
  9.   &lt;chr&gt; &lt;dbl&gt; &lt;dbl&gt;
  10. 1 A       0.4     1
  11. 2 B       0.6     2
  12. 3 C       0.8     5

答案2

得分: 3

尝试

  1. library(dplyr) # 版本 >= 1.10
  2. df %>%
  3.    reframe(pct = mean(value == value[class]), value = value[class], .by = group)
  • 输出
  1.    group pct value
  2. 1     A 0.4     1
  3. 2     B 0.6     2
  4. 3     C 0.8     5

或者使用 data.table 选项

  1. library(data.table)
  2. setDT(df)[df[(class)], .(value = i.value, pct = mean(value == i.value)), on = .(group), by = .EACHI]
  3.    group value pct
  4. 1:     A     1 0.4
  5. 2:     B     2 0.6
  6. 3:     C     5 0.8
英文:

Try

  1. library(dplyr)#version &gt;= 1.10
  2. df %&gt;%
  3.    reframe(pct = mean(value == value[class]), value = value[class], .by = group)

-output

  1.    group pct value
  2. 1     A 0.4     1
  3. 2     B 0.6     2
  4. 3     C 0.8     5

Or with a data.table option

  1. library(data.table)
  2. setDT(df)[df[(class)], .(value = i.value,
  3.   pct = mean(value == i.value)), on = .(group), by = .EACHI]
  4.    group value pct
  5. 1:     A     1 0.4
  6. 2:     B     2 0.6
  7. 3:     C     5 0.8
  10. </details>
  14. # 答案3
  15. **得分**: 3
  17. 使用`ave`和`subset`的基本R选项：
  19. ```R
  20. subset(
  21.   transform(
  22.     df,
  23.     pct = ave(ave(class, group, value) > 0, group)
  24.   ), 
  25.   class
  26. )

得到结果：

  1.    group class value pct
  2. 2      A  TRUE     1 0.4
  3. 8      B  TRUE     2 0.6
  4. 11     C  TRUE     5 0.8

请注意，我只翻译了代码部分，没有包括附加的信息或回答。

英文:

A base R option with ave + subset

  1. subset(
  2.   transform(
  3.     df,
  4.     pct = ave(ave(class, group, value) &gt; 0, group)
  5.   ), 
  6.   class
  7. )

gives

  1.    group class value pct
  2. 2      A  TRUE     1 0.4
  3. 8      B  TRUE     2 0.6
  4. 11     C  TRUE     5 0.8

huangapple
  • 本文由 发表于 2023年3月31日 23:31:33
  • 转载请务必保留本文链接：https://go.coder-hub.com/75900281.html
  • dataframe
  • dplyr
  • r
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定