提出了一种新的神经元模型，其数据和权值用余数制表示，称之为余数制神经元．它的处理能力等效于一些多层感知机模型．同时提出用中国余数定理重构信号，等效于神经元之间的非线性连接．将此模型用于分类，其单层网络可用于解高阶和边界模糊问题 A new neuron model is proposed, whose data and weights are represented by the remainder system, which is called the remainder neuron. Its processing power is equivalent to some multilayer perceptron model. At the same time, it is proposed to reconstruct the signal by the Chinese remainder theorem, which is equivalent to the nonlinear connection between neurons. Using this model for classification, its single-layer network can be used to solve high order and boundary ambiguity problems