Reduction of Input Data

A research on a technique for reducing input data tables, based on clustering cases, attributes, and attribute values in such a way that operations are preserved was continued by Soe Than, as his Ph.D. research. The reduced table is a result of applying three partitions to the sets of all cases, attributes, and attribute values, respectively. The corresponding system was implemented by Susan Siegler and then re-implemented by Soe Than. Often, rules induced from reduced tables are simpler than rules induced from the original tables.

Return to the Main Page