Data Discretization with CAIM (Class-attribute Interdependence Maximization)ΒΆ

Non-Causal Step

In this step, CAIM algorithm, which maximizes the class-attribute interdependence and generates a (possibly) minimal number of discrete intervals, is used to discretize the data. A priori knowledge on the number of intervals is not required.

Input Parameters

  1. Input data

Output Parameters

  1. Discretized data

Workflow

../../../_images/workflow41.svg

Algorithm

Class-attribute Interdependence Maximization (CAIM)

References

  • L.A. Kurgan, K.J. Cios, CAIM Discretization Algorithm, IEEE Transactions on Knowledge and Data Engineering, vol. 16(2), pp. 145-153, 2004.