Overcomplete Blind Source Separation for Time-Series Processes

Ray-Bing Chen ； Shih-Wei Chiang

《中國統計學報》 44卷4期 (2006/12) Pp. 342-363

https://doi.org/10.29973/JCSA.200612.0001

並列摘要

Overcomplete blind source separation problems are considered here. The overcomplete situation is that the number of sources is greater than the number of receivers. In this work, sources are natural sounds, and time-series processes are used to characterized the statistical regularities of these natural sounds. The null-space algorithm is applied here to accomplish the source separation. We demonstrate several real examples based on two simple time-series processes. Finally we compare the separation results according to the different model assumptions.

並列關鍵字

Autoregressive process ； Autoregressive and moving average process ； Null-space algorithm ； Overcomplete independent component analysis

參考文獻

Bell, A. J.,Sejnowski, T. J.(1995).An in formation maximization approach to blind separation and blind deconvolution.Neural Computation.7(6),1129-1159.

Bickel, P. J.,Levina, E.(2004).Some theory for Fisher`s linear discriminat function `naive Bayes`, and some alternative when there are many more variables than observations.Bernoulli.10(6),989-1010.

Box, G. E. P.,Jenkins, G. M.,Reinsel, G. C.(1994).Time Series Analysis: Forecasting and Control.Upper Saddle River, NJ:Prentice Hall.

Chen, R.-B.(2003).A Null-space Algorithm for Overcomplete Independent Component Analysis.Department of Statistics, University of California.

Chen, R.-B.,Wu, Y. N.(2002).A Null-space Representation for Overcomplete Independent Component Analysis. 2002 Proceedings of American Statistical Association, Statistical Computing Section [CD-ROM].Alexandria, VA:American Statistical Association.