Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this community
I wonder how Orange decides which attribute is declared as a string? I have few discrete attributes in Excel, which I can't use as Orange does not recognize them as discrete. Have tried export excel to csv format as well, no difference. Where in Orange I can enforce change from meta to discrete attribute?
Many thanks in advance, Marko
As far as I know, Orange supports advanced annotation with a three-line header in both CSV and Excel files. So if you have a discrete attribute that gets marked as string, you should annote it as discrete on the second header line, or prefix the attribute name on the first line with D#.
When no advanced annotation is provided, it seems Orange's heuristic treats attributes containing general strings (as opposed to numbers) as discrete only if they contain fewer than approximately $\sqrt{|data|}$ unique values.
