×

Blind signal separation of mixtures of chaotic processes: a comparison between independent component analysis and state space modeling. (English) Zbl 1277.94007

Summary: We perform a systematic comparison between different algorithms for solving the blind signal separation problem. In particular, we compare five well-known algorithms for independent component analysis (ICA) with a recently proposed algorithm based on linear state space modeling (IC-LSS). The comparison is based on simulated mixtures of six source signals, five of which are generated by nonlinear deterministic processes evolving on chaotic attractors. The quality of the reconstructed sources is quantified by two measures, one based on a distance measure implemented by a Frobenius norm, and one based on residual mutual information. We find that the IC-LSS modeling algorithm offers several advantages over the ICA algorithms: it succeeds in unmixing Gaussian sources, on short time series it performs, on average, better than static ICA algorithms, it does not try to remove coincidental dependencies resulting from finite data set size, and it shows the potential to reconstruct the sources even in the case of noninvertible mixing. As expected, for the case of non-Gaussian sources, invertible mixing and sufficient time series length, the ICA algorithms typically outperform IC-LSS modeling.

MSC:

94A12 Signal theory (characterization, reconstruction, filtering, etc.)
37M10 Time series analysis of dynamical systems
94A13 Detection theory in information and communication theory
68T05 Learning and adaptive systems in artificial intelligence
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] DOI: 10.1016/0005-1098(80)90078-3 · Zbl 0441.93036 · doi:10.1016/0005-1098(80)90078-3
[2] DOI: 10.1162/089976699300016458 · doi:10.1162/089976699300016458
[3] DOI: 10.1162/neco.1998.10.6.1373 · doi:10.1162/neco.1998.10.6.1373
[4] DOI: 10.1162/neco.1995.7.6.1129 · Zbl 05479838 · doi:10.1162/neco.1995.7.6.1129
[5] DOI: 10.1109/78.554307 · doi:10.1109/78.554307
[6] DOI: 10.1162/089976699300016863 · doi:10.1162/089976699300016863
[7] DOI: 10.1109/TSMCB.2003.811132 · doi:10.1109/TSMCB.2003.811132
[8] DOI: 10.1016/S0925-2312(01)00699-3 · Zbl 1006.68814 · doi:10.1016/S0925-2312(01)00699-3
[9] DOI: 10.1016/j.sigpro.2005.12.006 · Zbl 1172.93407 · doi:10.1016/j.sigpro.2005.12.006
[10] DOI: 10.1016/0165-1684(94)90029-9 · Zbl 0791.62004 · doi:10.1016/0165-1684(94)90029-9
[11] DOI: 10.1016/0167-2789(93)90103-8 · Zbl 0803.58038 · doi:10.1016/0167-2789(93)90103-8
[12] DOI: 10.1162/neco.2007.19.4.934 · Zbl 1210.94033 · doi:10.1162/neco.2007.19.4.934
[13] DOI: 10.1080/01621459.1981.10477720 · doi:10.1080/01621459.1981.10477720
[14] DOI: 10.1007/s10955-006-9131-x · Zbl 1108.62090 · doi:10.1007/s10955-006-9131-x
[15] Galka A., Kalman Filtering, ed. Gomez, J. (2011)
[16] DOI: 10.1007/s11538-010-9563-y · Zbl 1221.92020 · doi:10.1007/s11538-010-9563-y
[17] Geweke J., Latent Variables in Socio-Economic Models, eds. Aigner, D. & Goldberger, A. (North-Holland, Amsterdam) pp 365– (1977)
[18] DOI: 10.1111/1467-9892.00319 · Zbl 1036.62074 · doi:10.1111/1467-9892.00319
[19] DOI: 10.1016/j.csda.2004.12.012 · Zbl 1445.62224 · doi:10.1016/j.csda.2004.12.012
[20] DOI: 10.1016/j.neuroimage.2008.07.032 · doi:10.1016/j.neuroimage.2008.07.032
[21] DOI: 10.1109/TBME.2010.2046325 · doi:10.1109/TBME.2010.2046325
[22] DOI: 10.1109/72.761722 · doi:10.1109/72.761722
[23] DOI: 10.1016/j.sigpro.2005.02.003 · Zbl 1148.94400 · doi:10.1016/j.sigpro.2005.02.003
[24] DOI: 10.1007/BF02293808 · Zbl 0392.62043 · doi:10.1007/BF02293808
[25] DOI: 10.1016/j.jedc.2011.03.009 · Zbl 1217.91153 · doi:10.1016/j.jedc.2011.03.009
[26] DOI: 10.1016/0165-1684(91)90079-X · Zbl 0729.73650 · doi:10.1016/0165-1684(91)90079-X
[27] DOI: 10.1109/TAC.1968.1099025 · doi:10.1109/TAC.1968.1099025
[28] DOI: 10.1115/1.3662552 · doi:10.1115/1.3662552
[29] DOI: 10.1016/0304-4076(95)01774-7 · Zbl 0864.62059 · doi:10.1016/0304-4076(95)01774-7
[30] DOI: 10.1103/PhysRevE.69.066138 · doi:10.1103/PhysRevE.69.066138
[31] DOI: 10.1177/0049124103262065 · doi:10.1177/0049124103262065
[32] DOI: 10.1162/089976699300016719 · doi:10.1162/089976699300016719
[33] Mehra R. K., Ann. Econ. Social Measur. 3 pp 21– (1974)
[34] DOI: 10.1007/BF02294246 · Zbl 0603.62099 · doi:10.1007/BF02294246
[35] DOI: 10.1007/BF02294064 · Zbl 0621.62107 · doi:10.1007/BF02294064
[36] Ozaki T., Stat. Sin. 2 pp 113– (1992)
[37] DOI: 10.1239/jap/1085496593 · Zbl 1008.62086 · doi:10.1239/jap/1085496593
[38] DOI: 10.2307/1913421 · Zbl 0291.62124 · doi:10.2307/1913421
[39] DOI: 10.1093/imamci/3.2-3.185 · Zbl 0626.62088 · doi:10.1093/imamci/3.2-3.185
[40] DOI: 10.2514/3.3166 · doi:10.2514/3.3166
[41] DOI: 10.1109/MSPEC.1970.5213471 · doi:10.1109/MSPEC.1970.5213471
[42] DOI: 10.1103/PhysRevE.70.066123 · doi:10.1103/PhysRevE.70.066123
[43] DOI: 10.1016/0167-2789(92)90102-S · Zbl 1194.37144 · doi:10.1016/0167-2789(92)90102-S
[44] DOI: 10.1109/TCSI.2010.2052485 · doi:10.1109/TCSI.2010.2052485
[45] DOI: 10.1109/31.76486 · Zbl 0737.93016 · doi:10.1109/31.76486
[46] DOI: 10.1007/s11063-005-1000-0 · Zbl 05028187 · doi:10.1007/s11063-005-1000-0
[47] DOI: 10.1016/0304-4076(83)90066-0 · Zbl 0534.62083 · doi:10.1016/0304-4076(83)90066-0
[48] DOI: 10.1093/biomet/84.2.489 · Zbl 0882.62088 · doi:10.1093/biomet/84.2.489
[49] DOI: 10.1080/01621459.1999.10473861 · doi:10.1080/01621459.1999.10473861
[50] DOI: 10.1109/78.847796 · Zbl 0982.94004 · doi:10.1109/78.847796
[51] Zhang L., J. Sign. Process. 4 pp 111– (2000)
[52] Zhang K., J. Mach. Learn. Res.: Workshop and Conf. Proc. 20 pp 113– (2011)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.