Predictions from a latent class panel model

Predictions from a latent class panel model

Usage

# S3 method for class 'panfmm'
predict(
  object,
  newdata = NULL,
  type = c("response", "class", "prior", "posterior", "components"),
  weights = c("prior", "posterior"),
  ...
)

Arguments

object

a "panfmm" fit

newdata

optional data.frame; by default the estimation sample (in its original row order). Offsets are included in the class means: the estimation-sample offsets without newdata, the offset() terms of the formula evaluated in newdata otherwise (an offset argument of panfmm() cannot be evaluated in newdata and is not applied, with a warning)

type

"response" for the marginal mean sum_c w_ic mu_c(x_it); "class" for the modal class of each observation's individual (an integer, the class with the largest posterior probability in the estimation sample, or the largest prior probability, from the concomitant variables when there are some, with newdata); "posterior" / "prior" for the matrix of class probabilities of each observation's individual; "components" for the matrix of class-specific means mu_c(x_it) (one column per class). In panglm 1.5.0 and earlier, type = "class" returned the class-specific means now given by "components"; a message recalls this once per session.

weights

for type = "response": "prior" (the class probabilities, from the concomitant variables when there are some) or "posterior" (the in-sample posterior probabilities; only without newdata)

...

unused

Value

a numeric vector ("response"), an integer vector of class labels 1, ..., k ("class"), or a matrix with one column per class ("posterior", "prior", "components")

See also

Other panfmm methods: plot.panfmm(), summary.panfmm(), vcov.panfmm()

Examples

set.seed(3)
d <- data.frame(id = rep(1:150, each = 3), t = rep(1:3, 150), x = rnorm(450))
d$y <- rpois(450, exp(ifelse(rep(rbinom(150, 1, 0.4), each = 3) == 1, 1.5, 0) + 0.3 * d$x))
fit <- panfmm(y ~ x, data = d, index = c("id", "t"), nstart = 2, seed = 1)
head(predict(fit))
#> [1] 1.690137 2.067845 2.442469 1.596154 2.396391 2.279392
table(predict(fit, type = "class"))              # modal class per row
#> 
#>   1   2 
#> 282 168 
head(predict(fit, type = "posterior"))           # class probabilities
#>         class1     class2
#> [1,] 0.9112466 0.08875336
#> [2,] 0.9112466 0.08875336
#> [3,] 0.9112466 0.08875336
#> [4,] 0.9229608 0.07703924
#> [5,] 0.9229608 0.07703924
#> [6,] 0.9229608 0.07703924
predict(fit, newdata = data.frame(x = c(-1, 0, 1)), type = "components")
#>         class1   class2
#> [1,] 0.7443682 3.187864
#> [2,] 0.9513565 4.399313
#> [3,] 1.2159025 6.071134