Saturday, 7 September 2013

Are there any 'advanced' discretization techniques for WEKA?

Are there any 'advanced' discretization techniques for WEKA?

Using WEKA, I am running some analyses on the influence of discretization
on different algorithms and data sets.
To date, I am aware of the following three discretization techniques
implemented in WEKA:
Equal-width and equal-frequency binning (unsupervised>attributes>discretize)
Equal-frequency with sqr(k) bins, where k is the number of distinct values
for the attribute considered (unsupervised>attributes>PKIdiscretize)
Entropy-based discretization (supervised>attributes>discretize)
Apart from these three, I am not aware of any other implementation for WEKA.
Are there any other ones I didn't notice or additional packages available?

No comments:

Post a Comment