中文信息自动分类系统及其神经网络优化算法

对两种在我校曾用过的中文自动分类方法进行 了介绍．在对这二种自动分类方法的改进基础上，提出了一种全新的基于神经网络的自动分 类优化算法．这种方法运用人工神经网络理论，并结合专家的职业经验，模仿手工分类的思 维过程，对过去已研制成的两个系统进行了改进和优化．这种方法的设计思想、体系结构、 理论性优化在这里进行了重点介绍．

NEURAL NETWORK OPTIMIZING ALGORITHM OF CHINESE INFORMATION AUTO-CLASSIFICATION

According to two old methods of autoclassification of Chinese document, we provide an optimizing method based on the neural network algorithm. This method uses the theory of neural network, combines with the professional experiences of experts, imitates the process of manual classification, and optimizes the last two classification methods. The theory, architecture, and optimization of the new method are showed in the paper.