Global Variables

The program uses the following global variables:

N length of the data set
K number of states
m dimension of data vectors, y_t
nu recombination rate
TransP[1:K,1:K] matrix of transition probabilities [S_(t+1),S_t]
P0[1:K]' initial probabilities
D[1:m,1:N] training data: D=(y_1,...,y_N)
Dgen[1:m,1:N] test data: D=(y_1,...,y_N)
ScaleFact[1:N] scaling factors, computed in the forward path
Alpha[1:K,1:N] array of alpha_t(s_t), computed in the forward path
Beta[1:K,1:N] array of beta_t(s_t), computed in the backward path
PSD[1:K,1:N] array of posterior probabilities P(s_t|D)
PSSD[1:K,1:K,1:(N-1)] array of posterior probabilities P(s_t,s_(t+1)|D)
EmissionP[1:K,1:N] array of emission probabilities for all states
HMM_O Object: Emission model of the HMM
alpha_prior, beta_prior prior values of the hyperparameters
alpha_post, beta_post posterior values of the hyperparameters


Last modified: Mon May 22 13:38:34 BST