- unirandp.c Generate a Uniform Random Probability (0.0<p<1.0)
- invnorm.c Compute the Inverse Normal CDF (Peter J. Acklam)
- normcdf.c Compute the univariate Normal CDF (Abramowiz and Stegun 1964)
- normpdf.c Compute the univariate Normal PDF
- normsamp.c Generate Normal Samples (Box-Muller)
- bvmix.c Bivariate Mixtures (Shotaro Akaho 2001)
- bvmodnorm.c Normal module for bvmix
- bvmoduni.c Uniform module for bvmix
- bvncdf.c Compute the Bivariate Normal CDF integral (Cumulative Distribution Function) (Hull, 1993)
- bvnpdf.c Numerical computation of the Bivariate Normal PDF (Probability Distribution Funcion)
- bvnsamp.c Generate correlated BiVariate Normal samples
- mvmix.c Multivariate Mixtures
- mvmodnorm.c Normal module for mvmix
- mvmoduni.c Uniform module for mvmix
- mvnsamp.c Generate correlated Multivariate Normal Samples (with Cholesky decomposition)
- mvutils.c Matrix and vector utilities for Multivariate statistics
- prob_test.c Tests of the above
Example of use: finding 3 gaussian kernels
After a dozen of iterations...
kernel 0 weight 0.008567
--
kernel 1 weight 0.329915
mu:
| 314.051009|
| 451.104374|
| 357.568529|
covariance matrix:
| 4536.292629 121.080205 70.424394 |
| 121.080205 374.984352 -14.281386 |
| 70.424394 -14.281386 2685.141184 |
--
kernel 2 weight 0.286790
mu:
| 95.959635|
| 89.039549|
| 123.249103|
covariance matrix:
| 722.296765 175.224044 697.816596 |
| 175.224044 462.309905 636.365172 |
| 697.816596 636.365172 4512.438561 |
--
kernel 3 weight 0.374728
mu:
| 278.446285|
| 54.096309|
| 291.849748|
covariance matrix:
| 3496.551162 -638.048528 4556.283832 |
| -638.048528 235.603541 -780.323339 |
| 4556.283832 -780.323339 6769.880339 |
--
likelihood -15.774254