The paper offer a new heuristic attribute reduction algorithm based on conditional granularity entropy, though running an example, we show that this algorithm is effective.

提出了一种基于条件粒度熵的属性约简的启发式算法，通过例子分析，表明该算法是有效的。