After making a deep study of the problems of the existing inductive learning algorithms based on discernibility matrix a摘要深入分析了基于分辨矩阵和传统决策矩阵的归纳学习算法中存在的问题，提出了一种新的决策矩阵及基于它的归纳学习和动态数据环境下的递增式学习方法。
Among the decomposed synthesis reasoning sources, the field intensities were defined by dependency and classification me在分解得到的综合源中，基于正域的概念，用依赖度和分类质量方法定义了场强，根据属性出现频率，釆用差别矩阵定义了场强，并结合信息论，通过互信息、条件熵和互信息增益率定义了场强。
Based on substitution of the indiscernibility relation by a dominance relation and a confident threshold value a set, an摘要通过用优势关系代替不可分辨关系，并设定置信阈值α，提出了一种可以从偏好多属性决策表中发现偏好概率规则集的扩展粗糙集模型。
In this paper,a novel rough set model based on concordance relation is presented,which relaxing the indiscernibility and首先在传统的粗糙集模型基础上,通过松弛对象之间的不可分辨和相容性条件,给出了一种基于和谐关系的扩展粗糙集模型;在新模型的基础上,提出了一种新的层次聚类方法;给出一个具体的实例对新的模型加以解释说明。
The problem of calculating the core attributes of a dicision table is studied,a new discernibility matrix and the comput论文对目前求核方法存在的问题进行了分析,提出了一种新的可区分矩阵与求核方法,并证明了方法的正确性,该方法适用于任何决策表(相容的或不相容的)。