Problem 9.1 - purpose of parameter "target" in entropyGain()?

[system]

Disclaimer: Dieser Thread wurde aus dem alten Forum importiert. Daher werden eventuell nicht alle Formatierungen richtig angezeigt. Der ursprüngliche Thread beginnt im zweiten Post dieses Threads.

[reptor]

Problem 9.1 - purpose of parameter „target“ in entropyGain()?

public abstract class DecisionTable {
    // TODO should return Gain(attr) (for "target" as query-attribute) for the node in the tree where all "assumes" is satisfied
    protected abstract double entropyGain(Attribute attr, List<AttVal> assumes, Attribute target);
}

I don’t get why entropyGain() has a third parameter called „target“, and the explanation (for „target“ as query-attribute) is not helpful either, as I would rather think of the first parameter „attr“ as the query-attribute.

Maybe target is the „querying“ attribute? But how does that matter when „assumes“ already contains all the information necessary to construct the proper subset of examples?

Can someone who understood this please provide a better explanation?

EDIT: d’oh, I got it, target == tennis attribute, splitting the examples into yes/no ones