Formatting data for use in Apriori

[iank_23]

Hello, new to knime. I have a database table that I would like to perform Itemset Frequency Analysis (Apriori) on, as such I am trying to create a transaction list. I’ve filtered the table into the two needed columns; a Customer Id and a Product ID. Using the Group By node, the table is in a format where one Customer entry with numerous products, essesntially a list of transactions made by every unique customer.

I tried to feed this into the WEKA Apriori node, but apparently it doesn’t work with strings, how can I format it to work properly?

[elsamuel]

Welcome to the forum @iank_23.

I tried to feed this into the WEKA Apriori node, but apparently it doesn’t work with strings, how can I format it to work properly?

As far as I can tell, the Apriori node does work with strings.

Using the Group By node, the table is in a format where one Customer entry with numerous products, essesntially a list of transactions made by every unique customer.

I believe that this is the problem. I think that the data should be formatted like this:

And the Apriori node gives an output that looks like this (you’d have to play around with the settings):

Alternatively, you could use the Association Rule Learner node. You’d convert your concatenated list to a collection:

then apply the Association Rule Learner node:

[iank_23]

Thanks for the reply. Could you please explain how to achieve the formatting of the first image?

[mlauber71]

@iank_23 in addition I can offer this collection of rule algorithms

Rule Induction with Weka Rule Nodes and Yacaree Associator – mlauber71

Need idea in cluster analysis problem

I have tried the few lines but they do not make that much sense and for the Weka nodes I had in mind they do not have the right format. So I used an older example I had from a Kaggle DS. It is just for illustration the values do not make much sense with regards to a sequence. These basic differences Weka HotSpot can deal with strings and numbers and would accept a Target (in your case if you want to differentiate between Errors and Non-Errors). This might potentially handle your duration value…

[iank_23]

Thanks, but for some reason I still cant get any of these to work. I think my list should be read as nominal values but all I get is this:

ERROR Apriori (3.7)        5:14       Execute failed: Weka associator cannot work with given data. Reason: weka.associations.Apriori: Cannot handle string attributes!

Any help appreciated.

[iank_23]

Could you share the workflow you used if possible?

[elsamuel]

This is what I put together:

Itemset mining – elsamuel