作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (18): 171-173. doi: 10.3969/j.issn.1000-3428.2012.18.046

• 人工智能及识别技术 • 上一篇    下一篇

用于能耗数据分析的改进并行BP算法

周 媛 1,宋海涛 2,蒋砚军 1   

  1. (1. 北京邮电大学计算机学院,北京100876;2. 新华通讯社通信技术局,北京100803)
  • 收稿日期:2011-11-16 修回日期:2012-01-09 出版日期:2012-09-20 发布日期:2012-09-18
  • 作者简介:周 媛(1987-),女,硕士研究生,主研方向:数据挖掘;宋海涛,高级工程师;蒋砚军,副教授、硕士
  • 基金资助:
    国家自然科学基金资助项目(61074128)

Improved Parallel BP Algorithm for Energy Consumption Data Analysis

ZHOU Yuan 1, SONG Hai-tao 2, JIANG Yan-Jun 1   

  1. (1. College of Computer Science, Beijing University of Posts and Telecommunications, Beijing 100876, China; 2. Bureau of Communication Technology, Xinhua News Agency, Beijing 100803, China)
  • Received:2011-11-16 Revised:2012-01-09 Online:2012-09-20 Published:2012-09-18

摘要: 基于Map Reduce框架的传统BP神经网络算法收敛缓慢,训练易陷入局部极小点,使迭代次数过多,极大浪费资源。为此,提出并实现改进的并行BP算法,采用动态调节学习率、动量因子调整权重修正值,提升BP网络并行训练效率,利用预处理数据和最大分类概率增强分类的准确性。实验结果表明,改进的并行算法能提高分类准确率,缩短近17/18的训练时间。

关键词: 神经网络, 改进反向传播算法, Map Reduce架构, 并行, 学习率, 动量因子

Abstract: The traditional algorithm of BP neural network based on Map Reduce framework has been researched. But the convergence speed of traditional BP algorithm is quite slow. The training is easily into local minimum spot. The iterations of processing training data are quite frequency, a great waste of resources. The proposed design thus parallels the improved BP neural network based on Map Reduce. The adjustment of self-adaptive learning rate and momentum factors are implemented in this design. The proposed design also uses maximum classification probability factor and preprocess method to enhance the correctness. Experimental result shows that improved parallel algorithm costs 17 times shorter than the primary. The accuracy is also improved further.

Key words: neural network, improved Back Propagation(BP) algorithm, Map Reduce architecture, parallel, learning rate, momentum factor

中图分类号: