Strange Ranking of independent attributes via GainRatio

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Strange Ranking of independent attributes via GainRatio

Gerhard Marini
Dear all,

applying the GainRatio Ranker to a set of _independent_ attributes I got  
the following strange ranking values:

0.7638
0.7638
0.7638
0.7638
0.7638
0.7638
0.7638
0.7638
0.62824
0.62824
0.62824
0.53774
0.531
0.52825
0.52825
0.52825
0.52825
0.52825
0.52825
0.52825
0.52825
0.52825
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
0.44913
....  an so on

What could be the reason for this result? I definitely know that the  
attributes are independent, so I am rather puzzled ...

Thanks for your help in advance,

Gerhard

_______________________________________________
Wekalist mailing list
[hidden email]
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist