基于Kohonen和BP神经网络的文本学习算法

摘要

This paper introduces a text-learning algorithm based on Kohonen and BP neural network on Internet.It adopts vector space model to encode the text,and uses the Self-Organizing feature of Kohonen neural network and the nonlinear feature of BP neural network.After trained,the algorithm can present the rating which measure the closeness to the user＇s interests.So it can be used to perform the tasks,such as information filtering and intelligent browsing on Internet,etc.%介绍了基于Kohonen和BP神经网络结合的Internet网上文本学习算法。它采用向量空间模型对文本进行编码,利用Kohonen网络的自组织特性和BP网络的非线性特性进行学习。经过训练,算法能够有效地对输入文本进行判断, 给出一个评价等级, 标识出文本和用户兴趣的相关程度, 从而为基于Internet的信息过滤、智能浏览等处理提供基础。