基于粗糙集的综合推理模型

为了拓展综合推理的涵盖范围，建立了基于粗糙集理论的综合推理模型.将综合推理中的综合源和场引入到粗糙集理论的决策系统中. 通过对原始决策系统的分解得到综合源，分解而成的每个决策系统构成一个综合推理的源.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，采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.分析结果表明，该模型可完成属性约简过程.实现了对粗糙集理论和综合推理理论的融合和成功扩展.

Synthesis reasoning model based on rough set theory

For generalizing the content of synthesis reasoning, the model of synthesis reasoning based on rough set theory was proposed. The concepts of synthesis source and field in synthesis reasoning were introduced into decision systems in rough set theory. The synthesis sources were decomposed from the initial decision system, and a synthesis reasoning source was constructed from each decomposed decision system. Among the decomposed synthesis reasoning sources, the field intensities were defined by dependency and classification methods based on the concept of positive region, by discernibility matrix method based on the rate of appearance, and by mutual information, condition entropy and mutual information gain rate methods based on information theory. Analysis results show that the model can implement attribute reduction. Rough set theory and synthesis reasoning theory can be integrated and generalized successfully.