Intelligent Miner® provides
different clustering algorithms.
These clustering algorithms
are:
- Distribution-based (demographic) clustering
- Enhanced BIRCH clustering
- Center-based neural network (Kohonen) clustering
With these clustering algorithms, you can specify the following
optional parameters for fields in a logical data specification:
- The lower bounds and the upper bounds for numerical fields by
using the method DM_setFldOutlLim
- A name mapping for categorical fields by using the methods DM_setFldNmp and DM_addNmp
- The number of discrete bins to be used to discretize continuous
numeric input fields by using the method setFldNumBins
With the Clustering algorithms, you cannot specify taxonomies for
a categorical fields.