神经网络是通过模拟生物神经系统的结构功能而构成的一种信息处理系

统。经过40余年的发展，神经网络已经成为机器学习、模式识别、信号处理等

信息科学领域中一类十分重要的方法。

本文首先概括地总结了神经网络的发展历程。在三个不同阶段中，神经网

络的发展具有不同的特点。到20世纪90年代，神经网络成为信息科学中一个

重要的热点研究领域。文中还讨论了神经网络的特点、常见的分类方法、学习

类型和学习算法等内容。

-Through the simulation of biological neural network structure and function of the nervous system consisting of an information processing system. After 40 years of development, the neural network has become a machine learning, pattern recognition, signal processing, information science in a class of very important ways. This article first briefly summarizes the development process of the neural network. In three different stages in the development of neural networks with different characteristics. To the 1990s, information science, neural networks become an important hot research field. The paper also discusses the characteristics of neural networks, a common classification methods, learning styles and learning algorithm and so on.