Adaptive mixture methods based on Bregman divergences
Digital Signal Processing: A Review Journal
86 - 97
Item Usage Stats
We investigate adaptive mixture methods that linearly combine outputs of m constituent filters running in parallel to model a desired signal. We use Bregman divergences and obtain certain multiplicative updates to train the linear combination weights under an affine constraint or without any constraints. We use unnormalized relative entropy and relative entropy to define two different Bregman divergences that produce an unnormalized exponentiated gradient update and a normalized exponentiated gradient update on the mixture weights, respectively. We then carry out the mean and the mean-square transient analysis of these adaptive algorithms when they are used to combine outputs of m constituent filters. We illustrate the accuracy of our results and demonstrate the effectiveness of these updates for sparse mixture systems.
Published Version (Please cite this version)http://dx.doi.org/10.1016/j.dsp.2012.09.006
Showing items related by title, author, creator and subject.
Zengin, V.; Gökmen, A.; Dinçer, S.; Süzer, Ş. (Elsevier, 1995)Optical emission in 1:1 binary gaseous mixtures under high voltage displays a varying character depending on pressure, applied voltage polarity, and chemical nature of the mixture. Under negative polarity, in pure N2 and ...
Tanatar, Bilal; Erdemir, E. (Elsevier Science B.V., 2003)The density profiles of bosonic and fermionic components in a system of trapped two-dimensional (2D) boson-fermion (BF) mixture are studied. We employ the variational approach to minimize the total energy functional of the ...
Taşgin, M. E.; Subaşi, A. L.; Oktel, M. Ö.; Tanatar, Bilal (Springer, 2005)We consider a trapped system of atomic boson-fermion mixture with a quantized vortex. We investigate the density profiles of bosonic and fermionic components as functions of the boson-boson and boson-fermion short-range ...