Association rules with more than one item in the consequent

KNIME Analytics Platform
1

Hi,

do you have any plan to support that?

Thanks,

Alfredo

2

Hi Alfroc,

no plans sorry. I did never see a algorithm for it, but I guess it is NP complete? 

Do you know any good one?

Cheers, Iris 

3

Hi Alfroc,

this could be easily done in R using arules library with these dependencies using the common apriori algorithm:

library(caret)
library(arules)
library(arulesCBA)
frequent.rules <- apriori(here_your_data,parameter = list(supp = 0.1,conf = 0.6),
              appearance = list(rhs=c("your_column_value=example_value","another_column_value=example_value"), 
                                   default="lhs"))
rules.sorted <- sort(frequent.rules, by="confidence")

inspect(rules.sorted)
```

So you can get all the association rule combinations with 2 items as consequent (rhs) ordered by confidence value. 

-Giulio

4

Hi Giulio, Apriori only creates rules with one item in the RHS (Consequent) as well!

Cheers,

Alfredo

5

Hi Iris,

please take a look to this: https://www.it.uu.se/edu/course/homepage/infoutv/ht08/vldb94_rj.pdf

Cheers,

Alfredo