Split module fails with 4 columns

Question

Hello.
I try to make a predictive model for a movie rating prediction with own datasets. I have prepared and joined data and now I want to split the data into training and testing datasets. I have 4 columns (title, rank, number of votes and year) in the table and
the split module fails with warning "Error 0022: Number of selected columns in input dataset does not equal to 4." Why?
Thank you.

Answer 1

Hi Lukas,
The recommender splitter works under the assumption that it's splitting recommendation data, which we assume to consist either of user-item pairs or user-item-rating triples. Hence the limit of 2 or 3 columns, which is imposed to try and avoid confusion
with feature-type data.
-Y-