• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2008, Vol. 30 ›› Issue (4): 50-52.

• 论文 • 上一篇    下一篇

一种改进的BP神经网络手写体数字识别方法

王燕   

  • 出版日期:2008-04-01 发布日期:2010-05-19

  • Online:2008-04-01 Published:2010-05-19

摘要:

本文介绍了利用BP神经网络来识别手写体数字的原理,分析了传统BP算法收敛速度较慢的原因,提出了一种提高网络收敛速度的改进算法。实验数据表明,该改进算法比传统的算法在网络收敛速度上提高了九倍左右。

Abstract:

This paper presents the theory to recognize handwritten digits by the BP neural network, and an improved algorithm to accelerate the running speed of   the network based on the analysis of the traditional BP algorithm' s running speed. The experimental results show the running speed of the improved alg  gorithm is nine times that of the traditional algorithm.

Key words: neural network, BP algorithm, handwritten digit, feature extraction