RapidMiner

RapidMiner

order rules by support value (APRIORI algorithm)

order rules by support value (APRIORI algorithm)

Hi,

when using the APRIORI algorithm for itemset extraction, rules are by default ranked by confidence.
How can I rank the rules by support value? By changing the T parameter? Which value should T have?

Many thanks in advance.
Geert
3 REPLIES
Elite

Re: order rules by support value (APRIORI algorithm)

Hi Geert,

as far as I can see, you are using a WEKA operator. Unfortunately we cannot give support for their operators, mainly because we don't know them either. I would suggest using the FPGrowth implementation of rapid miner. If you have enough memory, this algorithm is a much faster alternative to apriori and the results are shown in a much more userfriendly way. You can apply the association rule generator afterwards and you will be able to sort the table in one click after any of the measures.


Greetings,
  Sebastian

Re: order rules by support value (APRIORI algorithm)

Hi,

and how can I rank the rules by support value? Only confidence,lift, conviction, ps, gain and laplace can be selected in my gui window?

Thanks a lot,
Geert
RMStaff

Re: order rules by support value (APRIORI algorithm)

Hi,

simply click on the column header "Support" in the table should sort the itemsets or rules according to the selected column. This is true for almost all tables presented in RapidMiner.

Cheers,
Ingo