An Improved Independent Component Analysis with Reference

Yan Fei Jia; Xiao Dong Yang; Li Yue Xu; Li Quan Zhao

doi:10.4028/www.scientific.net/AMM.667.64

Paper Titles

Design and Implementation in Project Based Case Learning System for Education in Software Engineering
p.41

Research of Information Integration for the Grid Panoramic Information Platform Based on SOA
p.50

Research on the Construction of Library Knowledge Service Platform
p.55

Research on Performance Evaluation of Finance Transportation Projects Based on Fuzzy Neural Network
p.60

An Improved Independent Component Analysis with Reference
p.64

The Research of BIM Theory Applied in Decision Making of Commercial Building Green Retrofit
p.68

Study of Crew Scheduling System for High-Speed Railway
p.72

Key Technologies for High-Speed Railway Timetable Parameter Design Based on Train Operation
p.76

Study on High-Speed Railway Train Scheduling Based on Integration of Management and Control
p.81

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vol. 667An Improved Independent Component Analysis with...

An Improved Independent Component Analysis with Reference

Abstract:

Independent component analysis with reference is a general framework to incorporate a priori information of interesting source signal into the cost function as constrained terms to form an augmented Lagrange function, and utilizes Newton method to optimize the cost function. It can extract any interesting source signal without extracting all source signals comparing with the traditional Independent component analysis method. In this paper, to accelerate the convergence speed of the Independent component analysis with reference, two improved algorithms are presented. The new algorithms, firstly whiten the observed signals to avoid matrix inverse operation to reduce algorithm complexity, secondly use improved Newton method with fast convergence speed to optimize cost function,in the end deduce the improved Independent component analysis with reference algorithms. Simulation result demonstrates the new algorithms have faster convergence speed with smaller error compared with the original method.

You might also be interested in these eBooks

Advances in Computers, Electronics and Mechatronics

View Preview

Info:

Periodical:

Applied Mechanics and Materials (Volume 667)

Pages:

64-67

DOI:

https://doi.org/10.4028/www.scientific.net/AMM.667.64

Citation:

Cite this paper

Online since:

October 2014

Authors:

Yan Fei Jia, Xiao Dong Yang, Li Yue Xu, Li Quan Zhao*

Keywords:

Convergence Speed, Independent Component Analysis with Reference, Interesting Signal, Newton Method

Export:

RIS, BibTeX

Price:

Permissions CCC:

Request Permissions

Permissions PLS:

Request Permissions

Сopyright:

Citation:

* - Corresponding Author

References

[1] W Lu, J Rajapakse. 3rd international conference on independent component Analysis and blind source separation, San Diego, California, (2001). p.120.

Google Scholar

[2] W Lu, J Rajapakse. advances in neural information processing systems. (2000) , p.570.

Google Scholar

[3] L Qinghua, Z rongrui Y fuliang. Information sciences. vol. 177(2007) , p.1265.

Google Scholar

[4] Zhang Shoucheng, Liu Yongkai. Computer Engineering and applications. vol(48)(2012), p.130.

Google Scholar

[5] Li Changli, L Guisheng, S Yuli. Digital signal processing. vol. 20 (2010), p.575.

Google Scholar

[6] W Lu, J Rajapakse. IEEE Transactions on neural networks. vol. 16 (2005) , p.203.

Google Scholar

[7] S Zhangli. IEEE transactions on information forensics and security. vol. 6 (2011), p.360.

Google Scholar

[8] L Breuer, M Axer. Journal of Neuroscience methods. vol. 220 (2013), p.30.

Google Scholar

[9] W Zhiyang, C Jin, D Guangming. Mechanical systems and signal processing. vol. 25 (2011), p.2501.

Google Scholar

[10] W Xiao-feng. mathematics practice and theory, vol. 40 (2010) , p.216.

Google Scholar

[11] S Qifang, L Xiwen. Journal of Taizhou University, vol. 30(2008), p.1.

Google Scholar