gmix_step: EM Baggenstoss: gmix_step
| gmix_step | R Documentation |
EM Baggenstoss: gmix_step
Description
Subroutine to update Gaussian mixture (5 Operations): 1. E-M algorithm (gmix_step) E-M algorithm (expectation maximization algorithm) See the matlab documentation for more information.
Usage
gmix_step(Parm, Data, Bias=0, DataWTS=NULL)
Arguments
Parm |
Nested list with parameters for GMM. Features carrying permanent values. Features$name [1:d] String vector with feature names. Features$min_std [1:NMODE] Vector of covariance constraints. Modes carrying modifyable values. Modes$cholesky_covar [d*NMODE, d] Numerical matrix with NMODE many square matrices stacked vertically with the covariance matrix. Modes$mean [1:NMODE, d] Numerical matrix with nmode different means and d feature dimensions. Modes$weight [1, 1:NMODE] Numerical matrix with weights for each mean. |
Data |
[1:n,1:d] Numerical matrix with normalized data. N samples with DIM feature dimensions. |
Bias |
Optional: Default=0. Binary value: Covariance constraint method. Choose: 1=BIAS, 0=CONSTRAINT |
DataWTS |
Optional: Numerical Vector, which allows individually weighting input data. Default=matrix(1,N,1). |
Value
List with two elements:
Parm |
Nested list with parameters for GMM. Parm$features carrying permanent values. Parm$features$name [1:d] String vector with feature names. Parm$features$min_std [1:NMODE] Vector of covariance constraints. Parm$modes carrying modifyable values. Parm$modes$cholesky_covar [d*NMODE, d] Numerical matrix with NMODE many square matrices stacked vertically with the covariance matrix. Parm$modes$mean [1:NMODE, d] Numerical matrix with nmode different means and d feature dimensions. Parm$modes$weight [1, 1:NMODE] Numerical matrix with weights for each mean. |
Q |
Numerical value: Total log-likelihood output (weighted by data_wts). |
Author(s)
Quirin Stier
References
Baggenstoss, Paul M., and T. E. Luginbuhl.: An EM algorithm for joint model estimation. IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258), Phoenix, AZ, USA, 1999, pp. 1825-1828 vol.4, IEEE, doi:10.1109/ICASSP.1999.758276, 1999.
