Up Next
Go up to 1 Decision-tree Learning
Go forward to Solution to part(b)

Solution to part (a)

What is a good attribute to split on first? Explain why.

a is a good attribute to split on, as when a is true, all of the examples agree on the value of attribute d, and when a is false the examples are skewed.

b is also a good attribute to split on, as it divides the examples into the same ratios as a.

c is a worse attribute to split on, as knowing the value of c doesn't help us much in determining the value of d.


Computational Intelligence online material, ©David Poole, Alan Mackworth and Randy Goebel, 1998

Up Next