The following node is available in the Open Source KNIME predictive analytics and data mining platform version 2.7.1. Discover over 1000 other nodes, as well as enterprise functionality at http://knime.com.

CAIM Binner

This node implements the CAIM binning (discretization) algorithm according to Kurgan and Cios (2004) URL:http://citeseer.ist.psu.edu/kurgan04caim.html. The binning (discretization) is performed with respect to a selected class column. CAIM creates all possible binning boundaries and chooses those that minimize the class interdependancy measure. To reduce the runtime, this implementation creates only those boundaries where the value and the class changes. The algorithm finds a minimum number of bins (guided by the number of possible class values) and labels them "Interval_X". Only columns compatible with double values are binned and the column's type of the output table is changed to "String".

Dialog Options

Class Column
The class column. According to this column the binning is optimized.
Column selection
Allows to include those columns which should be included in the discretization. Just the included columns are discretized and changed to "String" type.
General Node Settings
To increase performance, select the memory policy 'Keep all in memory' at the PREVIOUS node (if possible) from the General Node Settings Tab.

Ports

Input Ports
0 The data table to bin (discretize).
Output Ports
0 The binned data table.
1 The model representing the binning. Contains the intervals for each bin of each column.

Views

Binning Model
The view shows the column's binning scheme. For each column a ruler is displayed on which the bin boundaries are marked.
This node is contained in KNIME Base Nodes provided by KNIME GmbH, Konstanz, Germany.