R - arules 将关联规则封装为数据框

Question

我在 R 中使用 arules 包生成了关联规则。 已为 6 列/字段生成规则。 我想要的是一个由 6 列组成的数据框，这些列应该由关联规则填充。

例如：-

1. lhs rhs 支持信心提升
2. {性别=M,婚姻状况=Y,有工作吗？=Y} {Loan=Y} 0.7 0.8 1.9
3. {性别=F,婚姻状况=Y,有工作吗？=Y} {Loan=Y} 0.6 0.7 1.1
4. {Sex=M,有工作吗？=N} {Loan=N} 0.3 0.9 14.0

这应该以这种方式放入数据框中。

• Sex MaritalStatus 有工作吗？贷款支持信心提升
• M Y Y Y 0.7 0.8 1.9
• F Y Y Y 0.6 0.7 1.1
• M - N N 0.3 0.9 14

Answer 1

这需要一些编码和理解 R 和 arules 中使用的数据结构。这里有一些代码（希望）能做你想做的事。

library(arules)

# create some data
dat <- data.frame(
   Sex = c("M", "F", "M"), 
   Status = c("Y", "Y", "N"), 
   Job = c("Y", "Y", "N"),
   Loan = c("Y", "Y", "N")
   )

trans <- as(dat, "transactions")
itemInfo(trans)

#     labels variables levels
# 1    Sex=F       Sex      F
# 2    Sex=M       Sex      M
# 3 Status=N    Status      N
# 4 Status=Y    Status      Y
# 5    Job=N       Job      N
# 6    Job=Y       Job      Y
# 7   Loan=N      Loan      N
# 8   Loan=Y      Loan      Y

# arulesCBA can mine classification rules (CARs) with items for the 
# class variable in the RHS.
library(arulesCBA)
rules <- mineCARs(Loan ~ ., trans, parameter = list(supp = 1/3, conf = 0))
inspect(head(rules))

#     lhs           rhs      support   confidence lift count
# [1] {}         => {Loan=N} 0.3333333 0.3333333  1.0  1    
# [2] {}         => {Loan=Y} 0.6666667 0.6666667  1.0  2    
# [3] {Sex=F}    => {Loan=Y} 0.3333333 1.0000000  1.5  1    
# [4] {Status=N} => {Loan=N} 0.3333333 1.0000000  3.0  1    
# [5] {Job=N}    => {Loan=N} 0.3333333 1.0000000  3.0  1    
# [6] {Sex=M}    => {Loan=N} 0.3333333 0.5000000  1.5  1    

# rules store information about how the items relate to the original variables
ii <- itemInfo(rules)
ii

#     labels variables levels
# 1    Sex=F       Sex      F
# 2    Sex=M       Sex      M
# 3 Status=N    Status      N
# 4 Status=Y    Status      Y
# 5    Job=N       Job      N
# 6    Job=Y       Job      Y
# 7   Loan=N      Loan      N
# 8   Loan=Y      Loan      Y

# start with translating the rules into a logical matrix
m <- as(items(rules), "matrix")
head(m)

#      Sex=F Sex=M Status=N Status=Y Job=N Job=Y Loan=N Loan=Y
# [1,] FALSE FALSE    FALSE    FALSE FALSE FALSE   TRUE  FALSE
# [2,] FALSE FALSE    FALSE    FALSE FALSE FALSE  FALSE   TRUE
# [3,]  TRUE FALSE    FALSE    FALSE FALSE FALSE  FALSE   TRUE
# [4,] FALSE FALSE     TRUE    FALSE FALSE FALSE   TRUE  FALSE
# [5,] FALSE FALSE    FALSE    FALSE  TRUE FALSE   TRUE  FALSE
# [6,] FALSE  TRUE    FALSE    FALSE FALSE FALSE   TRUE  FALSE

# do some R tricks to create the data.frame
df <- do.call(cbind, 
   lapply(unique(ii$variables), FUN = function(var) {
   cols <- which(ii$variables == var)
   df <- data.frame(factor(apply(t(m[,cols])*(1:length(cols)), MARGIN = 2, max), 
     levels = 1:length(cols), 
     labels = ii$levels[cols]))
   colnames(df) <- var
   df
   }))

# add quality measures
df <- cbind(df, quality(rules))
head(df)

#    Sex Status  Job Loan   support confidence lift count
# 1 <NA>   <NA> <NA>    N 0.3333333  0.3333333  1.0     1
# 2 <NA>   <NA> <NA>    Y 0.6666667  0.6666667  1.0     2
# 3    F   <NA> <NA>    Y 0.3333333  1.0000000  1.5     1
# 4 <NA>      N <NA>    N 0.3333333  1.0000000  3.0     1
# 5 <NA>   <NA>    N    N 0.3333333  1.0000000  3.0     1
# 6    M   <NA> <NA>    N 0.3333333  0.5000000  1.5     1