论文部分内容阅读
针对动态信息系统中新增和退出对象集的情况,在信息观下提出了一种基于条件熵的核属性计算方法。先分别讨论仅有对象集进入、仅有对象集退出以及以上两种情形同时存在下条件熵的变化机制。由条件熵变化机制,通过构造支持度矩阵和增量矩阵方式将新增或者退出的对象集一并参与比较计算,快速求得条件熵,进而获得核属性。最后,由实例分析验证该方法的有效性。
Aiming at the new and exited set of objects in the dynamic information system, a new method of kernel attribute calculation based on conditional entropy is proposed. First, we discuss the mechanism of change of conditional entropy when only the object set enters, only the set of objects exits and the above two conditions exist simultaneously. According to the mechanism of change of conditional entropy, the new or exited set of objects are all involved in comparative computation by constructing support matrix and incremental matrix, and the conditional entropy is obtained quickly to obtain the core attributes. Finally, the validity of this method is verified by an example analysis.