General problem with every weka nodes (Apriori, PredictiveApriori, Tertius... )

KNIME Analytics Platform
#1

Hi collegues,

this is a more general and extended problem that I'm facing with every kind of nodes belonging to weka. I have already talked about this problem in a previous topic here, but i didn't received a clear clarification of how to solve it.

Every kind of input data (without missing vaues) that I'm trying to use is not compatible and the knime console log returns this error:

Execute failed: Weka associator cannot work with given data. Reason: weka.associations.Tertius: Cannot handle string attributes!

Of course if I try to convert the input data as integer or in every other type of data format, the expected error is always the same with the specific data format input that I have tried to use.

Can someone explain me with a simple example workflow which kind of data the weka nodes exectly require? 

Since for example in the Borgelt node is require a list of aggregated values. But in this case? The data need to be grouped or not? I just need to know how to correctly prepare it.

Thanks in advice.

-Giulio

#2

The Weka nodes always use all incoming columns. If a Weka algorithm cannot handle string columns you have to remove them out beforehand.
 

#3

Hi @thor,

can you please show me a simple working knime example with a few of these weka nodes that I have quote?

So I can understand what I need to pass as input data in these situations.

Thanks in advice.

-Giulio