DR by BINNING

DR by BINNING

1. WHAT?

   • A data discretisation method
     • Aimed at reducing the number of values assumed by an attribute.
       • HOW TO REDUCE?
         • By transforming continuous attributes into categorical attributes characterized by a limited number of distinct values.
           • EG
             • the weekly spending of a mobile phone customer is a continuous attribute.
               • With binning, the continuous value of weekly spending can be binned into five categories as follows:
                 • In this case, the binning process reduces the number of distinct values for the attribute of weekly spending of a mobile phone customer.

2. EG OF BINNING METHODS

   • • Subjective subdivision
     • Classes are defined based on the experience and judgment of domain experts.
   • • Automated subdivision
     • Subdivision into categories may be automatically determined based on equal size or equal width

3. IBM SPSS Modeler

   • Binning node allows the automatic conversions of numerical fields into categorical fields.
     • EG
       • an income field can be transformed into a new categorical field containing income groups of equal width

Imagens e arquivos do mapa