基于多模型的不等长序列数据关联算法
发布时间：2018-11-23 13:59
【摘要】：单模型在处理不等长序列数据关联时不能兼顾计算精度、复杂度和抗扰性,为此提出了基于多模型(MM)的不等长序列数据关联算法。将基于滑动窗口和动态时间弯曲(DTW)的不等长序列相似度度量模型作为MM的输入模型,以2种模型计算得到的时似变化比作为模型判断指标进行模型转换,实现了2种模型的优势互补,并得到模型的应用条件,最后输出MM作用后的不等长序列相似度,以此作为关联指标进行关联判定。仿真实验验证了MM关联算法在处理不等长序列数据关联的有效性,并对序列长度和突变率变化对关联效果的影响进行了分析。
[Abstract]:When dealing with unequal length sequence data association, a single model can not take into account computational accuracy, complexity and immunity. Therefore, an unequal length sequence data association algorithm based on multi-model (MM) is proposed. The similarity measurement model of unequal series based on sliding window and dynamic time bending (DTW) is used as the input model of MM, and the time likelihood ratio calculated by the two models is used as the model judgment index to transform the model. The advantages of the two models are complementary, and the application conditions of the models are obtained. Finally, the similarity of unequal length sequences after MM is output, which is used as an association index to determine the association. Simulation results show that the MM correlation algorithm is effective in dealing with data association of unequal length sequences, and the influence of the variation of sequence length and mutation rate on the correlation effect is analyzed.
【作者单位】： 海军航空工程学院电子信息工程系;91934部队;
【基金】：国家自然科学基金(61032001) 新世纪优秀人才支持计划(NCET-11-0872)~~
【分类号】：TP212;TP301.6
本文编号：2351819
[Abstract]:When dealing with unequal length sequence data association, a single model can not take into account computational accuracy, complexity and immunity. Therefore, an unequal length sequence data association algorithm based on multi-model (MM) is proposed. The similarity measurement model of unequal series based on sliding window and dynamic time bending (DTW) is used as the input model of MM, and the time likelihood ratio calculated by the two models is used as the model judgment index to transform the model. The advantages of the two models are complementary, and the application conditions of the models are obtained. Finally, the similarity of unequal length sequences after MM is output, which is used as an association index to determine the association. Simulation results show that the MM correlation algorithm is effective in dealing with data association of unequal length sequences, and the influence of the variation of sequence length and mutation rate on the correlation effect is analyzed.
【作者单位】： 海军航空工程学院电子信息工程系;91934部队;
【基金】：国家自然科学基金(61032001) 新世纪优秀人才支持计划(NCET-11-0872)~~
【分类号】：TP212;TP301.6
【相似文献】
相关期刊论文 前7条
1 朱晓钢;杨兵;许华杰;;支持无线传感器网络多目标跟踪的聚类数据关联算法研究[J];计算机科学;2012年S1期
2 叶西宁;顾幸生;潘泉;;数据关联算法性能评估的一种方法[J];火力与指挥控制;2006年05期
3 李辉;左现刚;张安;段航宇;;复杂环境下数据关联算法的研究现状及发展趋势[J];火力与指挥控制;2007年09期
4 徐洋;徐晖;罗少华;安玮;;基于随机有限集理论的多传感器目标联合检测跟踪算法[J];国防科技大学学报;2013年01期
5 ;电脑文摘[J];电脑开发与应用;1995年04期
6 李景熹;王树宗;王航宇;;密集杂波环境下的数据关联算法研究[J];系统仿真学报;2009年05期
7 黄伟平;徐毓;王杰;;应用回归分析的数据关联算法[J];西安交通大学学报;2011年08期
，本文编号：2351819
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2351819.html
