Abstract: In this paper, we give a more improved greedy algorithm for discretization of decision table. The importance of break point is decided amount of division a pair of example, and the importance of next break point is decided the choice of the break point when them are equally. The incertitude is removed in degree, and the algorithms are more perfect. The example of the end express that the algorithms may reduce the amount of break point and unknown the district.

Key words: decision table, algorithms, Rough sets, discretization

摘要： 对决策表属性离散化的贪心算法做了进一步改进。以分开实例对的多少确定断点的重要性，当出现同样重要的断点时，用下一个断点的重要性来决定选用哪一个断点，在一定程度上消除了不确定性，使算法更加完善。最后给出的实例表明，该算法能够减少断点数和未知区域。

关键词: 决策表, 算法, 粗糙集, 离散化