Categorization of transactions

Hi guys,

I have a database containing info on bank account transactions (account, date, amount, description/transfer reason) and I would like KNIME to automatically assign predefined categories to these transactions.

I have learning data available (i.e. transactions to which categories were assigned manually) and I would like KNIME to learn to assign categories based on "description/transfer reason" which usually is a string containing 1-3 words.

How would I need to approach this problem? Which machine learning methods and corresponding KNIME nodes would I need to use?

Thanks in advance for you support!

done

Hi done,

This sounds like a pretty straightforward document classification problem. You can find a lot of Text Processing examples in the KNIME node guide. In your case, this one should be particularly interesting: https://www.knime.org/nodeguide/other-analytics-types/text-processing/document-classification

Cheers,

Roland