The filter MathExpression can be found in this package:
It provides a powerful means of performing mathematical transformations of numeric attributes. The following operators are supported:
 +, -, *, /, (, ), 
 pow, log, abs, cos, exp, sqrt, tan, sin, ceil, floor, rint, 
The attribute value that is being processed, can be referenced as A.

Manual discretization

One can even use the filter for manually discretizing numeric attributes, if the other Discretize filters (supervised and unsupervised) cannot be used. This works thanks to the ifelse operator.

It is basically a two-step-process:
  1. run MathExpression to turn all the values into discrete ones
  2. run NumericToNominal to turn the numeric values then into nominal labels

Here's an example:
  • a dataset where the first attribute needs to be discretized into 3 bins
  • the bins need to be as follows
  • using MathExpression to create discrete values
 weka.filters.unsupervised.attribute.MathExpression \
   -E "ifelse(A>20, ifelse(A>80, 3, 2), 1)" \
   -V \
   -R 1
  • Note: -V -R 1 means we only want to transform the first attribute. Without -V all the numeric attributes would be transformed according to this expression.
  • this results in the following transformation
 (-inf-20.0] -> 1
 (20.0-80.0] -> 2
 (80.0-inf)  -> 3
  • using NumericToBinary to create a nominal attribute from the numeric one
 weka.filters.unsupervised.attribute.NumericToNominal \
   -R 1
  • optional: if one wants to rename those labels, one can use the class listed in the Rename Attribute Values article for that

Note: the "\" at the end of the lines tell a *nix bash to continue on the next line.