Hi,

It's a multi-label classification. The output here is the target labels :)

HTH.

Samer Sarsam, PhD.

Combine target columns into 1 and use them as classification labels.

Yes , You can use Artifical neural networks (ANN) for your classification problem .

Example input target pair for an multi class problem . You can use smilar data set for training and testing of ANN .

input target

[ x y z ] [ 0 0] // first class

[ a b c ] [ 0 1] // second class

[ k l m ] [ 1 0] // third class

[ u v y ] [ 1 1] // fourth class

I am not too sure of your choice of name. Yes, you may call it multi-label classification if you want as you have two columns that are labels/response variables/dependent variables. But actually when we are talking of binary classification or multi-classification, it has to do with the content (classes/categories) in your label column. so the question is how many classes or categories are the information in your column label variables. If you have two classes, it is binary classification and if you have more than two classes, it is multiclassification. Your first label column can be two classes and your second label column 4classes (who knows) and in that case, you may be dealing with the problem as a binary problem and also as a multi class problem. So, what i am saying in summary is that whether your classification problem is binary or multiclass is related to the categories in your label column and not the number of your label columns as you seem to infer. You may have 5-10 columns of dependent variables and will still be dealing with a binary class problem. I hope this helps.