一种格状分区模型的研究

Study of a Kind of Partition Model

中文关键词  分区模型  格状分区  递推分区  样本不均衡

英文关键词  partition model  grid partition  iterative partition  imbalanced training sets

中文摘要

      本文提出一种新的分区模型结构并给出辨识算法。与此同类型的其它模型结构相比，该模型算法简洁，占用内存少，鲁棒性好，并具有与输入分量尺度无关性和很好的收剑性。

英文摘要

      In this paper, a kind of partition model structure and its identification algorithm is presented. Compared with other structures of this group, this model has many good qualities such as simple algorithm, small consuming memory, good robustness and convergence. Furthermore, the identification result is ir-relavent to the scales of each input variables.