The paper put forward a method of discretization of continuous attributes based on discernibility matrix and revised the discrete result by average mutual information.

• 本文基于可辨识矩阵提出一种连续属性离散化的方法，并利用平均互信息量对离散化结果进行修正。