| prob_to_int {mHMMbayes} | R Documentation |
Transforming a set of probabilities to Multinomial logit regression intercepts
Description
prob_to_int transforms a set of state transition or categorical
emission observation probabilities to the corresponding Multinomial logit
regression intercepts. Note that the first category is assumed to be the
reference category, hence no intercept is returned for the first state or
category.
Usage
prob_to_int(prob_matrix)
Arguments
prob_matrix |
A matrix with number of states OR categories columns and
number of rows to be determined by the user, with rows summing to one. For
obtaining the set of Multinomial logit regression intercepts of the
complete transition probability matrix gamma or categorical emission
distribution matrix, the number of rows equals the number of states
|
Details
Designed to ease the specification of informative hyper-prior values for the
mean intercepts of the transition probability matrix gamma and categorical
emission distribution(s) of the multilevel hidden Markov model through the
functions prior_gamma and prior_emiss_cat. No
check is performed on correct specifications of the dimensions.
Value
prob_to_int returns a matrix containing Multinomial logit
regression intercepts, with the number of columns equal to (number of
states or categories - 1) and the number of rows equal to the number rows
specified in the input matrix. The first state / category is assumed to be
the reference category, hence no intercept is returned for this first
category.
See Also
int_to_prob for transforming a set of Multinomial
logit regression intercepts to a probabilities, prior_gamma
and prior_emiss_cat for specifying informative hyper-priors
for the the multilevel hidden Markov model and mHMM to fit a
multilevel hidden Markov model.
Examples
# example for transition probability matrix gamma with 3 states
m <- 3
gamma_prob <- matrix(c(0.6, 0.2, 0.2,
0.1, 0.8, 0.1,
0.1, 0.1, 0.8), ncol = m, nrow = m, byrow = TRUE)
gamma_int <- prob_to_int(gamma_prob)
gamma_int