One matrix of posterior latent-state probabilities for a fit whose
family carries a frmtmb_structure() with a latent_probs slot: the
probability of each state or class given everything the model
observed. Rows sum to one.
Value
A numeric matrix with one named column per latent state. Its
rows are observations for a family whose class belongs to the row
(lca(), a rowwise mixture()) and groups for one whose class
belongs to a group (mixture(groups = )).
Details
This is one entry point for every structured family, and the family
decides what it returns. mixture_probs(), hmm_probs() and
lca_probs() are the same computation under family-specific names
and checks; use whichever names what you fitted.
See also
frmtmb_structure(), mixture_probs(), hmm_probs(),
lca_probs(), hmm_viterbi() for the decoded path rather than the
probabilities
Examples
set.seed(1)
dd <- data.frame(y = c(rnorm(80, 0, 1), rnorm(80, 5, 1)))
fit <- frm(bf(y ~ 1) + mixture(gaussian(), gaussian()), data = dd)
head(latent_probs(fit))
#> class1 class2
#> [1,] 0.9999999 7.623679e-08
#> [2,] 0.9999948 5.179722e-06
#> [3,] 1.0000000 2.623868e-08
#> [4,] 0.9888568 1.114321e-02
#> [5,] 0.9999888 1.123676e-05
#> [6,] 1.0000000 2.833862e-08