点击次数：

论文类型：期刊论文

发表时间：2022-06-30

发表刊物：大连理工大学学报

卷号：51

期号：6

页面范围：933-936

ISSN号：1000-8608

摘要：The main function of fuzzy perceptron is to discriminate which
   categories the samples are in by weight learning. An algorithm for a
   recurrent fuzzy perceptron based on fuzzy logic is presented, and the
   network structure of the recurrent fuzzy perceptron is similar to
   traditional perceptron based on addition- production, and the dynamic
   recursion term is added. Initial weights of network are set to be
   constant zero, in the case where the dimension of the input vectors is
   two and the training examples are separable, its finite convergence is
   proved, i.e., the training procedure for the network weights will stop
   in finite steps, and when the dimension is greater than two, stronger
   conditions are needed to guarantee the finite convergence.

备注：新增回溯数据