Knowledge reduction algorithm based on mutual information
Huizhong Yang · Journal of Tsinghua University(Science and Technology) · 2007
One key purpose of rough set theory essence is knowledge reduction.This paper presents an algorithm for attribute reduction based on mutual information.Both the mutual information and the conditional information entropy between the selected attribute and the decision attribute are considered.An attribute importance measure method defined using information theory is used as the heuristic information in the algorithm.Tests show that this approach reduces the attributes in most cases.