问题
I am trying to find association rules using the apriori function from arules package in R.
rules <- apriori(data=data, parameter=list(supp=0.001,conf = 0.08),
appearance = list(default="lhs",rhs="YOGHURT"),
control = list(verbose=F))
rules <- sort(rules, decreasing=TRUE,by="confidence")
inspect(rules[1:3])
lhs rhs support confidence lift
1. {A,B} {C} 0.04 0.96 0.25
2. {C,A} {D} 0.05 0.95 0.26
3. {B,D} {A,C} 0.03 0.93 0.24
With the code showed above I got some association rules saved in the variable "rules" ordered by confidence in a decreasing way. But I would like to order these rules by confidence and by lift at the same time. I tried this but I got an error:
rules <- sort(rules, decreasing=TRUE,by=c("confidence","lift"))
Error in .subset2(x, i, exact = exact) : subscript out of bounds
Is there a way to sort rules by confidence and lift at the same time?
回答1:
Assuming you got
library(arules)
data("Adult")
rules <- apriori(Adult, parameter = list(supp = 0.5, conf = 0.9, target = "rules"))
then you could try
df <- as(rules, "data.frame")
df[order(df$lift, df$confidence), ]
回答2:
I have not thought about this. You can copy & paste the following code into your R session after you load arules.
setMethod("sort", signature(x = "associations"),
function (x, decreasing = TRUE, na.last = NA, by = "support", ...) {
q <- quality(x)
q <- q[, pmatch(by, colnames(q)), drop = FALSE]
if(is.null(q)) stop("Unknown interest measure to sort by.")
if(length(x) == 0) return(x)
x[do.call(order, c(q, list(na.last = na.last, decreasing = decreasing)))]
})
Now your original code should work.
> data("Adult")
> rules <- apriori(Adult, parameter = list(supp = 0.5, conf = 0.9, target = "rules"))
> inspect(head(sort(rules, by=c("supp", "conf"))))
lhs rhs support confidence lift
1 {} => {capital-loss=None} 0.9532779 0.9532779 1.0000000
2 {} => {capital-gain=None} 0.9173867 0.9173867 1.0000000
3 {capital-gain=None} => {capital-loss=None} 0.8706646 0.9490705 0.9955863
4 {capital-loss=None} => {capital-gain=None} 0.8706646 0.9133376 0.9955863
5 {native-country=United-States} => {capital-loss=None} 0.8548380 0.9525461 0.9992323
6 {native-country=United-States} => {capital-gain=None} 0.8219565 0.9159062 0.9983862
This will be part of the next release of arules.
来源:https://stackoverflow.com/questions/32106764/sorting-rules-by-lift-and-confidence