首页 | 官方网站   微博 | 高级检索  
     

条件过滤器的自适应排序调度策略
引用本文:刁静,赵荣彩,张铮,邰铭,刘勇.条件过滤器的自适应排序调度策略[J].信息工程大学学报,2011,12(4):441-446.
作者姓名:刁静  赵荣彩  张铮  邰铭  刘勇
作者单位:信息工程大学信息工程学院,河南郑州,450002
基金项目:国家863计划资助项目(2008AA011002)
摘    要:将非线性流形学习的思想引入到语音特征降维中,设计完成了局部投影(Locality Preser-ving Projections,LPP)算法,并针对该算法对降至的目的维度敏感的缺点提出了一种加权局部投影(Weighted Locality Preserving Projections,WLPP)算法。对LPP算法及WLPP算法进行了实验,实验证明LPP算法在对语音特征降维后也能有较高的准确率,同时,也证明了WLPP算法可以大大降低LPP算法对目的维度的敏感度。

关 键 词:线性降维  流形学习  局部保持映射  加权局部保持映射

Adaptive Ordering Scheduling of Selective-Filters
DIAO Jing,ZHAO Rong-cai,ZHANG Zheng,TAI Ming,LIU Yong.Adaptive Ordering Scheduling of Selective-Filters[J].Journal of Information Engineering University,2011,12(4):441-446.
Authors:DIAO Jing  ZHAO Rong-cai  ZHANG Zheng  TAI Ming  LIU Yong
Affiliation:CHEN Hong-chang,QI Xiao-qian,HUANG Hai(Institute of Information Engineering,Information Engineering University,Zhengzhou 450002,China)
Abstract:An algorithm based on nonlinear manifold is proposed in this paper.LPP algorithm is proved effective but very sensitive to embedding dimension,so WLPP algorithm is designed to overcome such a drawback.Several contrast experiments are designed and performed to evaluate the effectiveness of LPP and WLPP algorithms.The experimental data supports the method of using LPP algorithm and WLPP algorithm,and also shows that WLPP is superior to LPP in insensitivity to embedding dimension.
Keywords:linear dimensionality reduction  manifold learning  LPP  WLPP  
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《信息工程大学学报》浏览原始摘要信息
点击此处可从《信息工程大学学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号