中文摘要：

知识约简是粗糙集理论的重要研究内容之一。在不完备信息系统中，对联系度粗糙集模型的研究比较深入，但极少涉及知识约简问题。该文在进一步改进联系度粗糙集模型的基础上，研究该模型的知识约简。针对改进模型，提出多种知识约简定义，给出了这些约简之间的关系。通过定义初等分辨矩阵和属性重要度，介绍一种分配约简算法。实例分析说明了算法的有效性。

英文摘要：

Knowledge reduction is one of the most important issues in rough sets theory. In incomplete information systems, researchers study connection-degree-based rough sets model. However, knowledge reduction for this model has never been researched. In this paper, the connection-degree-based rough sets model is improved and knowledge reduction for this mode is examined. Several knowledge reductions are proposed and the relations between them are showed. Using elementai＇y discernibility matrix and the significance of attributes defined, a knowledge reduction algorithm for incomplete information systems is presented. Example analysis shows that this algorithm is valid.