在R中使用筛选器对项目进行随机化。

huangapple
go评论92阅读模式
英文:

Randomize items with a filter in R

问题

有没有一种方法可以基于筛选条件对DataFrame的行进行洗牌？例如，我有这个DataFrame：

  1. data=data.frame(id=c(3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26),
  2.                 name=c("restructuring","restructuring","restructuring","restructuring",
  3.                        "control","control","control","control","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","action filler","action filler","action filler","action filler","action filler","action filler","action filler","action filler")
  4.                )

其中，编号从3到6是'restructuring'，7-10是'control'，11-18是'clitic filler'，19-26是'action filler'，我希望name列在两个连续的行中不具有相同的值。

我尝试过：

  1. shuffled_data = data[sample(1:nrow(data)), ]

但这显然没有特定的条件。

英文:

is there a way to shuffle dataframe's rows based on a filter? For instance, I have this dataframe:

  1. data=data.frame(id=c(3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26),
  2.                 name=c("restructuring","restructuring","restructuring","restructuring",
  3.                        "control","control","control","control","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","clitic filler","action filler","action filler","action filler","action filler","action filler","action filler","action filler","action filler")
  4.                )

In which numbers from 3 to 6 are 'restructuring', 7-10 are 'control', 11-18 are 'clitic filler', 19-26 are 'action filler', and I'd like name column to not have the same value in 2 consecutive rows.

I tried:

shuffled_data= data[sample(1:nrow(data)), ]

But this obviously randomizes with no criteria

答案1

得分: 1

如果你的数据大小是这样的，我会进行一系列的随机洗牌，并找出符合你的条件的一个或多个：

  1. shuffle = function(data) {
  2.   data[sample(1:nrow(data)), ]
  3. }
  5. check = function(data) {
  6.   all(data$name[-1] != data$name[-nrow(data)])
  7. }
  9. set.seed(47)
  10. results = replicate(10000, shuffle(data), simplify = FALSE)
  11. results = results[sapply(results, check)]
  12. length(results)
  13. [1] 10
  14. ## 在10000次洗牌中，有10次符合你的条件
  16. ## 这是其中一种情况：
  17. results[[1]]
  18. #    id          name
  19. # 16 18 clitic filler
  20. # 21 23 action filler
  21. # 9  11 clitic filler
  22. # 20 22 action filler
  23. # 15 17 clitic filler
  24. # 24 26 action filler
  25. # 1   3 restructuring
  26. # 13 15 clitic filler
  27. # 7   9       control
  28. # 2   4 restructuring
  29. # 19 21 action filler
  30. # 6   8       control
  31. # 4   6 restructuring
  32. # 23 25 action filler
  33. # 3   5 restructuring
  34. # 22 24 action filler
  35. # 10 12 clitic filler
  36. # 18 20 action filler
  37. # 12 14 clitic filler
  38. # 5   7       control
  39. # 11 13 clitic filler
  40. # 8  10       control
  41. # 17 19 action filler
  42. # 14 16 clitic filler
英文:

If your data is about this size, I would do a bunch of random shuffles and find one(s) that meet your criteria:

  1. shuffle = function(data) {
  2.   data[sample(1:nrow(data)), ]
  3. }
  5. check = function(data) {
  6.   all(data$name[-1] != data$name[-nrow(data)])
  7. }
  9. set.seed(47)
  10. results = replicate(10000, shuffle(data), simplify = FALSE)
  11. results = results[sapply(results, check)]
  12. length(results)
  13. [1] 10
  14. ## 10 of the 10000 shuffles meet your criteria
  16. ## here's one:
  17. results[[1]]
  18. #    id          name
  19. # 16 18 clitic filler
  20. # 21 23 action filler
  21. # 9  11 clitic filler
  22. # 20 22 action filler
  23. # 15 17 clitic filler
  24. # 24 26 action filler
  25. # 1   3 restructuring
  26. # 13 15 clitic filler
  27. # 7   9       control
  28. # 2   4 restructuring
  29. # 19 21 action filler
  30. # 6   8       control
  31. # 4   6 restructuring
  32. # 23 25 action filler
  33. # 3   5 restructuring
  34. # 22 24 action filler
  35. # 10 12 clitic filler
  36. # 18 20 action filler
  37. # 12 14 clitic filler
  38. # 5   7       control
  39. # 11 13 clitic filler
  40. # 8  10       control
  41. # 17 19 action filler
  42. # 14 16 clitic filler

答案2

得分: 1

使用来自[这个答案][1]的函数，其中min.dist = 1：

  1. library(data.table)
  3. setorder(setDT(data), name)[
  4.   frank(prob_shuffler(cumsum(!duplicated(name)), 1L), ties.method = "random")
  5. ]

#> id name
#> 1: 3 restructuring
#> 2: 10 control
#> 3: 17 clitic filler
#> 4: 7 control
#> 5: 5 restructuring
#> 6: 13 clitic filler
#> 7: 20 action filler
#> 8: 11 clitic filler
#> 9: 9 control
#> 10: 24 action filler
#> 11: 16 clitic filler
#> 12: 25 action filler
#> 13: 14 clitic filler
#> 14: 19 action filler
#> 15: 4 restructuring
#> 16: 18 clitic filler
#> 17: 22 action filler
#> 18: 6 restructuring
#> 19: 23 action filler
#> 20: 15 clitic filler
#> 21: 8 control
#> 22: 21 action filler
#> 23: 12 clitic filler
#> 24: 26 action filler

  2.   [1]: https://stackoverflow.com/a/65013927/9463489
  4. <details>
  5. <summary>英文:</summary>
  7. Using the function from [this answer][1] with `min.dist = 1`:
  9.     library(data.table)
  10.     
  11.     setorder(setDT(data), name)[
  12.       frank(prob_shuffler(cumsum(!duplicated(name)), 1L), ties.method = &quot;random&quot;)
  13.     ]
  14.     #&gt;     id          name
  15.     #&gt;  1:  3 restructuring
  16.     #&gt;  2: 10       control
  17.     #&gt;  3: 17 clitic filler
  18.     #&gt;  4:  7       control
  19.     #&gt;  5:  5 restructuring
  20.     #&gt;  6: 13 clitic filler
  21.     #&gt;  7: 20 action filler
  22.     #&gt;  8: 11 clitic filler
  23.     #&gt;  9:  9       control
  24.     #&gt; 10: 24 action filler
  25.     #&gt; 11: 16 clitic filler
  26.     #&gt; 12: 25 action filler
  27.     #&gt; 13: 14 clitic filler
  28.     #&gt; 14: 19 action filler
  29.     #&gt; 15:  4 restructuring
  30.     #&gt; 16: 18 clitic filler
  31.     #&gt; 17: 22 action filler
  32.     #&gt; 18:  6 restructuring
  33.     #&gt; 19: 23 action filler
  34.     #&gt; 20: 15 clitic filler
  35.     #&gt; 21:  8       control
  36.     #&gt; 22: 21 action filler
  37.     #&gt; 23: 12 clitic filler
  38.     #&gt; 24: 26 action filler
  40.   [1]: https://stackoverflow.com/a/65013927/9463489
  42. </details>

huangapple
  • 本文由 发表于 2023年3月7日 01:58:23
  • 转载请务必保留本文链接：https://go.coder-hub.com/75654242.html
  • list
  • r
  • random
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