Converting object of class rules to data frame in R
I have an output of apriori function, which mines data and gives set of rules. I want to convert it to data frame for further processing.
The rules object looks like this:
> inspect(output)
lhs rhs support confidence lift
1 {curtosis=(846,1.27e+03]} => {skewness=(-0.254,419]} 0.2611233 0.8044944 2.418776
2 {variance=(892,1.34e+03]} => {notes.class=FALSE} 0.3231218 0.9888393 1.781470
3 {variance=(-0.336,446]} => {notes.class=TRUE} 0.2859227 0.8634361 1.940608
4 {skewness=(837,1.26e+03]} => {notes.class=FALSE} 0.2924872 0.8774617 1.580815
5 {entropy=(-0.155,386],
class=FALSE} => {skewness=(837,1.26e+03]} 0.1597374 0.9521739 2.856522
6 {variance=(-0.336,446],
curtosis=(846,1.27e+03]} => {skewness=(-0.254,419]} 0.1378556 0.8325991 2.503275
We can create rules object using data frame. Data frame looks like this:
> data
variance skewness curtosis entropy notes.class
1 (892,1.34e+03] (837,1.26e+03] (-0.268,424] (386,771] FALSE
2 (892,1.34e+03] (-0.254,419] (424,846] (771,1.16e+03] FALSE
3 (892,1.34e+03] (837,1.26e+03] (-0.268,424] (-0.155,386] FALSE
4 (446,892] (-0.254,419] (846,1.27e+03] (386,771] FALSE
Than we can get output variable using this:
> output <- apriori(data)
There was used arules package. dput(output) gives this:
new("rules"
, lhs = new("itemMatrix"
, data = new("ngCMatrix"
, i = c(8L, 2L, 0L, 5L, 9L, 12L, 0L, 8L, 0L, 3L, 0L, 8L, 8L, 13L, 8L,
10L, 3L, 10L, 8L, 11L, 8L, 13L, 3L, 12L, 2L, 5L, 2L, 6L, 2L,
5L, 2L, 6L, 2L, 10L, 2L, 7L, 2L, 11L, 0L, 3L, 0L, 10L, 0L, 7L,
11L, 13L, 5L, 6L, 6L, 12L, 5L, 10L, 1L, 5L, 4L, 6L, 6L, 13L,
0L, 3L, 8L, 0L, 8L, 13L, 3L, 8L, 13L, 0L, 3L, 13L, 2L, 5L, 6L,
2L, 5L, 12L, 2L, 6L, 12L)
, p = c(0L, 1L, 2L, 3L, 4L, 6L, 8L, 10L, 12L, 14L, 16L, 18L, 20L, 22L,
24L, 26L, 28L, 30L, 32L, 34L, 36L, 38L, 40L, 42L, 44L, 46L, 48L,
50L, 52L, 54L, 56L, 58L, 61L, 64L, 67L, 70L, 73L, 76L, 79L)
, Dim = c(14L, 38L)
, Dimnames = list(NULL, NULL)
, factors = list()
)
, itemInfo = structure(list(labels = structure(c("variance=(-0.336,446]",
"variance=(446,892]", "variance=(892,1.34e+03]", "skewness=(-0.254,419]",
"skewness=(419,837]", "skewness=(837,1.26e+03]", "curtosis=(-0.268,424]",
"curtosis=(424,846]", "curtosis=(846,1.27e+03]", "entropy=(-0.155,386]",
"entropy=(386,771]", "entropy=(771,1.16e+03]", "notes.class=FALSE",
"notes.class=TRUE"), class = "AsIs"), variables = structure(c(5L,
5L, 5L, 4L, 4L, 4L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L), .Label = c("curtosis",
"entropy", "notes.class", "skewness", "variance"), class = "factor"),
levels = structure(c(4L, 8L, 12L, 2L, 6L, 10L, 3L, 7L, 11L,
1L, 5L, 9L, 13L, 14L), .Label = c("(-0.155,386]", "(-0.254,419]",
"(-0.268,424]", "(-0.336,446]", "(386,771]", "(419,837]",
"(424,846]", "(446,892]", "(771,1.16e+03]", "(837,1.26e+03]",
"(846,1.27e+03]", "(892,1.34e+03]", "FALSE", "TRUE"), class = "factor")), .Names = c("labels",
"variables", "levels"), row.names = c(NA, -14L), class = "data.frame")
, itemsetInfo = structure(list(), .Names = character(0), row.names = integer(0), class = "data.frame")
)
, rhs = new("itemMatrix"
, data = new("ngCMatrix"
, i = c(3L, 12L, 13L, 12L, 5L, 3L, 8L, 13L, 0L, 3L, 8L, 3L, 3L, 8L,
6L, 5L, 12L, 12L, 12L, 12L, 12L, 13L, 13L, 13L, 3L, 12L, 5L,
12L, 12L, 13L, 4L, 13L, 3L, 0L, 8L, 12L, 6L, 5L)
, p = 0:38
, Dim = c(14L, 38L)
, Dimnames = list(NULL, NULL)
, factors = list()
)
, itemInfo = structure(list(labels = structure(c("variance=(-0.336,446]",
"variance=(446,892]", "variance=(892,1.34e+03]", "skewness=(-0.254,419]",
"skewness=(419,837]", "skewness=(837,1.26e+03]", "curtosis=(-0.268,424]",
"curtosis=(424,846]", "curtosis=(846,1.27e+03]", "entropy=(-0.155,386]",
"entropy=(386,771]", "entropy=(771,1.16e+03]", "notes.class=FALSE",
"notes.class=TRUE"), class = "AsIs"), variables = structure(c(5L,
5L, 5L, 4L, 4L, 4L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L), .Label = c("curtosis",
"entropy", "notes.class", "skewness", "variance"), class = "factor"),
levels = structure(c(4L, 8L, 12L, 2L, 6L, 10L, 3L, 7L, 11L,
1L, 5L, 9L, 13L, 14L), .Label = c("(-0.155,386]", "(-0.254,419]",
"(-0.268,424]", "(-0.336,446]", "(386,771]", "(419,837]",
"(424,846]", "(446,892]", "(771,1.16e+03]", "(837,1.26e+03]",
"(846,1.27e+03]", "(892,1.34e+03]", "FALSE", "TRUE"), class = "factor")), .Names = c("labels",
"variables", "levels"), row.names = c(NA, -14L), class = "data.frame")
, itemsetInfo = structure(list(), .Names = character(0), row.names = integer(0), class = "data.frame")
)
, quality = structure(list(support = c(0.261123267687819, 0.323121808898614,
0.285922684172137, 0.292487235594457, 0.159737417943107, 0.137855579868709,
0.137855579868709, 0.142231947483589, 0.142231947483589, 0.110138584974471,
0.110138584974471, 0.12399708242159, 0.153902261123268, 0.107221006564551,
0.13056163384391, 0.13056163384391, 0.150984682713348, 0.139314369073669,
0.100656455142232, 0.107221006564551, 0.154631655725748, 0.165572574762947,
0.112326768781911, 0.105762217359592, 0.12180889861415, 0.181619256017505,
0.181619256017505, 0.102844638949672, 0.105762217359592, 0.12837345003647,
0.12837345003647, 0.137855579868709, 0.137855579868709, 0.137855579868709,
0.137855579868709, 0.13056163384391, 0.13056163384391, 0.13056163384391
), confidence = c(0.804494382022472, 0.988839285714286, 0.863436123348018,
0.87746170678337, 0.952173913043478, 0.832599118942731, 0.832599118942731,
0.859030837004405, 0.898617511520737, 0.853107344632768, 0.915151515151515,
0.80188679245283, 0.972350230414747, 0.885542168674699, 0.864734299516908,
0.913265306122449, 1, 0.974489795918367, 1, 1, 0.990654205607477,
1, 0.980891719745223, 0.873493975903614, 0.814634146341463, 0.943181818181818,
0.950381679389313, 1, 0.92948717948718, 0.931216931216931, 0.897959183673469,
1, 0.969230769230769, 0.895734597156398, 0.832599118942731, 1,
0.864734299516908, 0.93717277486911), lift = c(2.41877587226493,
1.78146998779801, 1.94060807395104, 1.580814717477, 2.85652173913043,
2.50327498261071, 2.56515369004603, 1.93070701234925, 2.71366653809456,
2.56493458221826, 2.81948927477017, 2.41093594836147, 2.92344773223381,
2.72826587247868, 2.58853870008227, 2.73979591836735, 1.80157687253614,
1.75561827884899, 1.80157687253614, 1.80157687253614, 1.78473970550309,
2.24754098360656, 2.20459434060771, 1.96321350977681, 2.44926187419769,
1.69921455023295, 2.85114503816794, 1.80157687253614, 1.67454260588295,
2.09294821753838, 2.68799572230639, 2.24754098360656, 2.91406882591093,
2.70496064471679, 2.56515369004603, 1.80157687253614, 2.58853870008227,
2.81151832460733)), row.names = c(NA, 38L), .Names = c("support",
"confidence", "lift"), class = "data.frame")
, info = structure(list(data = data, ntransactions = 1371L, support = 0.1,
confidence = 0.8), .Names = c("data", "ntransactions", "support",
"confidence"))
)
Answer
We can't duplicate your data from your question (oh, you just added your data as I was typing this! Sorry!), so I'll use the example from the arules package:
library('arules');
data("Adult")
## Mine association rules.
rules <- apriori(Adult,
parameter = list(supp = 0.5, conf = 0.9,
target = "rules"))
Then I can duplicate the stuff output from inspect(rules):
> ruledf = data.frame(
lhs = labels(lhs(rules))$elements,
rhs = labels(rhs(rules))$elements,
rules@quality)
> head(ruledf)
lhs rhs support confidence lift
1 {} {capital-gain=None} 0.9173867 0.9173867 1.0000000
2 {} {capital-loss=None} 0.9532779 0.9532779 1.0000000
3 {hours-per-week=Full-time} {capital-gain=None} 0.5435895 0.9290688 1.0127342
4 {hours-per-week=Full-time} {capital-loss=None} 0.5606650 0.9582531 1.0052191
5 {sex=Male} {capital-gain=None} 0.6050735 0.9051455 0.9866565
6 {sex=Male} {capital-loss=None} 0.6331027 0.9470750 0.9934931
and do stuff like order by decreasing lift:
head(ruledf[order(-ruledf$lift),])
The help for the rules class: http://www.rdocumentation.org/packages/arules/functions/rules-class.html will tell you what you can get from your rules object - I just used that information to build a data frame. If its not exactly what you want, then cook one up using your own recipe!