I’m new to both Knime and association rule learning, so please forgive my very beginner question. I’m trying to bring a retail dataset into knime (it is retail.dat.gz found here http://fimi.uantwerpen.be/data/). It reads perfectly fine but when I try to connect it to the Association Rule Learner I get this error (screenshot attached):
" The dialog cannot be opened for the following reason:
No column in spec compatible to “BitVectorValue” or “CollectionDataValue”.
I literally have done nothing else, so I’m pretty sure there must be some simple setting that I need to change. Or maybe the data structure? Any help would be much appreciated.
Just a little step needed after you read the retail data, use the Cell Splitter node .
Settings: Enter a delimeter: whitespace
Output: As set (remove duplicates)
And don’t have your Minimum Support to high, otherwise no results will pop-up.
Hope this helps,
I’m sorry to bother about this, especially because I’m sure it’s another simple fix, but I do have one final question. Now, using the process Hans outlined above, my node is creating an empty data table. I lowered the support to .1 and still get no results. I’m playing with a bunch of the parameters but can’t seem to get results.
Also imported the data and set up a few association rule learners. You have to play with the confidence and the support (how many items have to be covered by the rule). And of course one would have to interpret what is meant by
Item set [48,41] leads to 39. And indeed out of the many sets, not too much rules result in the first place.
