Gmixt_simultanee calculates the cost matrix for a segmentation/clustering model

Gmixt_simultanee calculates the cost matrix for a segmentation/clustering model

Gmixt_simultanee(Don, lmin, phi)

Arguments

Don

the bivariate signal

lmin

the minimum size for a segment

phi

the parameters of the mixture

Value

a matrix G(i,j), the mixture density for segment between points (i+1) to j G(i,j) = $$\\sum_{p=1}^P \log (\\pi_p f(y^{ij};\\theta_p))$$

\sump=1 -> P log( \pip f( yij ; \thetap ) )

Rq: this density if factorized in order to avoid numerical zeros in the log