Association Rule in R: how to select rules based on item?

Question

I am using r arules package to generate rules around a transaction dataset. In the dataset, I have over 500 transactions with items such as apples, beer and so on.

I know how to generate the rules and sort them based on support or confidence, but if I want to only look at the rules that involve certain items, how should I do so? Like I only want rules that having apple in it.

Something like:

inspect(rules[keyword='apple'])

Answer 1

You can do that with subset.

inspect(subset(rules, subset = items %in% "apple"))

Since you do not provide your data, I will give a full example using data provided in the arules package.

library(arules)
data(Groceries)
rules <- apriori(Groceries, parameter = list(supp = 0.001, conf = 0.8))

Now pick out the rules that mention yogurt. There are too many to show the full result, so I just show the first three.

inspect(subset(rules, subset = items %in% "yogurt")[1:3])
    lhs                   rhs                    support confidence     lift count
[1] {yogurt,                                                                      
     cereals}          => {whole milk}       0.001728521  0.8095238 3.168192    17
[2] {yogurt,                                                                      
     rice}             => {other vegetables} 0.001931876  0.8260870 4.269346    19
[3] {other vegetables,                                                            
     yogurt,                                                                      
     specialty cheese} => {whole milk}       0.001321810  0.8125000 3.179840    13

None of these had yogurt on the rhs, so I also show rule 20 to show that it is catching yogurt there too.

inspect(subset(rules, subset = items %in% "yogurt")[20])
    lhs                                      rhs      support     confidence
[1] {other vegetables,butter milk,pastry} => {yogurt} 0.001220132 0.8       
    lift     count
[1] 5.734694 12