基于免疫应答的神经网络优化设计及其分类应用

Immune response based optimizing design for neural network and its classifying application

从信息存储记忆的角度观察，神经系统借助感官可存储和记忆外界信息，免疫系统则在抗原识别等方面表现出记忆功能．该文将生物免疫系统中的免疫应答识别原理应用到前馈神经网络中，利用免疫识别算法训练前馈神经网络，能够使网络优化过程趋于全局最优．将该算法对一组实验数据进行分类，仿真结果表明该方法具有良好的可行性和实用性．

From the view of information storing and memory, neural system can store and memorize outside information in virtue of sense, while immune system possesses memory through antigen recognition. This paper applies biological immune response recognition theory to feed-forward neural network, which is trained by immune recognition algorithm. The combination can optimize the network totally. Finally, a set of data is classified using the algorithm, and the simulation result demonstrates that it has better feasibility and practicability