西安电子科技大学学报 ›› 2016, Vol. 43 ›› Issue (4): 178-183.doi: 10.3969/j.issn.1001-2400.2016.04.031

• 研究论文 • 上一篇    下一篇

测量报告数据的谱分析压缩算法

程飞;刘凯;丁文文;时欢;张百戬   

  1. (西安电子科技大学 计算机学院,陕西 西安  710071)
  • 收稿日期:2015-10-19 出版日期:2016-08-20 发布日期:2016-10-12
  • 作者简介:程飞(1985-),男,西安电子科技大学博士研究生,E-mail: chengfei8582@163.com.
  • 基金资助:

    国家自然科学基金资助项目(61571345, 61550110247);中央高校基本科研业务费专项资金资助项目(BDY191420);安徽省高等学校自然科学研究资助项目(KJ2014B14)

Spectrum analysis compression algorithm of measure report data

CHENG Fei;LIU Kai;DING Wenwen;SHI Huan;ZHANG Baijian   

  1. (School of Computer Science and Technology, Xidian Univ., Xi'an  710071, China)
  • Received:2015-10-19 Online:2016-08-20 Published:2016-10-12

摘要:

针对网络带宽难以满足海量测量报告传输要求,定义了测量数据的谱,并提出了测量报告数据的谱分析压缩算法.该算法通过分析测量数据的谱,提出了对数据完成两次排序的预处理方案,减少了数据冗余的间隔距离,以期提高上下文的命中率.其次,该压缩算法构建了测量数据的多个上下文模型,并作为单层神经网络的输入结点.神经网络通过对每个上下文模型的预测概率线性组合,得到对下一比特的预测概率,并用真实的下一比特来调整对应输入结点的权重,提高匹配历史信息的概率.实验结果表明,在采用该压缩算法后,算法的运行时间明显减小,压缩率随被压缩数据量的增大而提高.与其他通用压缩算法相比,该算法能有效提高测量数据的压缩率,从而达到减少网络数据传输量的目的.

关键词: 测量报告, 数据压缩, 谱分析, 神经网络

Abstract:

To solve to the problem that with the increment of the number of mobile users, the network bandwidth cannot meet the mass transfer of the measure report. This paper defines the conception of the spectrum of the measure report and proposes a spectrum analysis compression algorithm for the measure report. The algorithm utilizes two step sorting in order to reduce the distance between similar contents according to the analysis of the spectrum of measure report data. Furthermore, this algorithm employs several context models, which are regarded as the input nodes of the neural network. The prediction probability is calculated by the linear combination of the probability of each context model, and the weight in each link is tuned by the next bit in order to increase the possibility of matching. Experimental results reveal that the algorithm not only decreases the compression consuming time, but also ascends the compression ratio with the increment of the size of compression data. Compared with other competitors, this algorithm can effectively increase the compression ratio of the measure report to gain a better transfer time.

Key words: measure report, data compression, spectrum analysis, neural network