论文部分内容阅读
将基于属性量-质特征转化的性质命题判断抽象为一个以定性基准为参量的动态定性映射,给出定性基准分别为一个区间、区间向量和 n 维超平行体网格的定义,讨论加权 w 对定性基准的几何意义,定性映射和人工神经元的关系,以及(性质判断的)定性映射转化为(模式)识别映射的机制和条件.指出.加权 w 的实质是对定性基准中的 n 维超平行体实施一个旋转变换.因人工神经元可看作是对 n 维超平行体的一个面实施加权 w 内积旋转变换后的形式,故是(加权 w 内积变换)定性映射的一个特例.当定性基准中 n 维超平行体的所有边长都趋近于零时,以(近)零边长的 n 维超平行体为基准的定性映射转化为一个向量的判断或搜索映射,并可对任一可转化为一个向量的模式进行识别.此外,本文还提出了属性量-质转化程度函数的概念和一种导致定性基准边界模糊化的机制,指出,若用某个模糊化的转化程度函数代替定性映射进行判断和识别,则得到一种模糊判断和识别,并以一系列实例证实了定性映射和转化程度函数方法的实用性.
We abstract the qualitative proposition based on attribute qualitative-qualitative feature transformation into a dynamic qualitative mapping based on qualitative datum. We give the definition of a grid with qualitative datum as an interval, interval vector and n-dimensional hyperparallel meshes, and discuss the relationship between weighted w This paper analyzes the geometric meaning of qualitative datum, the relationship between qualitative and artificial neurons, and the qualitative mapping of qualitative judgment into the mechanism and condition of (pattern) recognition mapping. The super-parallel body implements a rotation transformation, which is a special case of (qualitatively weighted w-inner product transform) qualitative mapping because the artificial neuron can be regarded as a form after the inner product of the inner product of the n-dimensional hyperparagraphs is rotated. When all sides of a n-dimensional hyper-parallel body in a qualitative datum approach zero, the qualitative mapping based on an n-dimensional hyperparagraph with (near) zero-side length transforms into a vector judgment or search mapping We can identify any pattern that can be transformed into a vector.In addition, we also propose the concept of attribute quantity-quality conversion degree function and a mechanism that leads to the blurring of the qualitative reference boundary. It is pointed out that if A fuzziness degree of transformation function is used instead of the qualitative mapping to judge and identify, then a kind of fuzzy judgment and identification is obtained. A series of examples are used to confirm the practicability of qualitative mapping and transformation degree function method.