Research Article

Gaussian Mixture Models Based on Principal Components and Applications

Table 2

Simulation case: parameter estimates of the bivariate GMM (first technique) (the log-likelihood, BIC, and number of iterations are provided).

Sample sizeModel parameters
First componentSecond component

500.53806−0.197440.461462.681851.356520.461940.22997−0.537502.353900.30614
Log-likelihood = −331.7415, BIC = 704.9295
Number of iterations: 70

1000.380401.07288−0.68292.330421.047020.61959−0.65870.419271.460890.73289
Log-likelihood = −332.917, BIC = 707.2805
Number of iterations: 96

5000. 65985−0.48985−0.20382.108791.007320.340150.95030.395322.292430.96703
Log-likelihood = −1678.241, BIC = 3379.508
Number of iterations: 77