Knime Queries - numeric/nominal values & defining columns

Hello this is my first post so I am sorry for being naive.

 
I have been using Weka and have been instructed to start using Knime but I have already reached a brick wall.
 
I have loaded a training set into Knime and I was first experimenting with using a decision tress, when I try to execute the decision tree node I get an error saying there is no numeric attributes. Which I do agree with tho when I am configuring the file reader node with the various attributes there are only 3 option, float, int or double. The first question is which of these do I set for a numeric or nominal value?
 
I have then continued to test the naive bayes classifier and again I get an error where I havnt defined the classification column, could anybody point me in the right direction to do this please. As I have been looking online with no luck.
 
I do apologise as these are simple questions but any help would be greatly appreciated.
 
Andy.

Hi Andy,

I guess the decision tree is complaining about no nominal attribute (not numerical)? Both that learner as well as the Naive Bayes are building classifiers, so they'd like to see at least one attribute with nominal (usually string) information as class column. You can configure the File Reader to read one (or more) of your columns as string (right click the column header in the preview!) or use a Binner node to convert one of the numerical columns into bins carrying a label.

Does this help?

Cheers, Michael

 

That really does help, thank you. I was curious how Knime reads a nominal value expecting it to be an int. and as you have expressed should have been a string. As to verififying the column  I'll give it a go after work.

Much appreciated the help.

Andrew