Extract the latent state estimates of a revision model

Accessor for the state paths of a fitted revision-nowcasting model. Provides programmatic access to the estimated states instead of reaching into the object with fit$states.

Usage

states(object, ...)

# S3 method for class 'kk_model'
states(object, filter = c("smoothed", "filtered", "all"), state = NULL, ...)

# S3 method for class 'jvn_model'
states(object, filter = c("smoothed", "filtered", "all"), state = NULL, ...)

Arguments

object

A fitted model object, such as kk_model or jvn_model.

...

Additional arguments passed to methods.

filter

Which state estimates to return: "smoothed" (default) uses the full sample, "filtered" uses information available up to each date, and "all" returns both.

state

Optional character vector of state names to keep. Defaults to all states.

Value

A tibble with columns time, state, estimate, lower, upper, filter and sample.

See also

Other revision nowcasting: coef.jvn_model(), coef.kk_model(), fitted.jvn_model(), fitted.kk_model(), jvn_nowcast(), kk_nowcast(), logLik.jvn_model(), logLik.kk_model(), nobs.jvn_model(), nobs.kk_model(), plot.jvn_model(), plot.kk_model(), predict.jvn_model(), predict.kk_model(), print.jvn_model(), print.kk_model(), residuals.jvn_model(), residuals.kk_model(), summary.jvn_model(), summary.kk_model(), vcov.jvn_model(), vcov.kk_model()

Examples

# \donttest{
gdp_growth <- dplyr::filter(
  tsbox::ts_pc(reviser::gdp),
  id == "EA",
  time >= min(pub_date),
  time <= as.Date("2020-01-01")
)
gdp_growth <- tidyr::drop_na(gdp_growth)
df <- get_nth_release(gdp_growth, n = 0:3)

fit <- jvn_nowcast(df = df, e = 4, ar_order = 2, include_noise = FALSE)
head(states(fit))
#> # A tibble: 6 × 7
#>   time       state      estimate     lower   upper filter   sample   
#>   <date>     <chr>         <dbl>     <dbl>   <dbl> <chr>    <chr>    
#> 1 2002-10-01 news_vint1   0.112   0.0141    0.210  smoothed in_sample
#> 2 2003-01-01 news_vint1   0.0212 -0.0766    0.119  smoothed in_sample
#> 3 2003-04-01 news_vint1   0.0983  0.000530  0.196  smoothed in_sample
#> 4 2003-07-01 news_vint1  -0.0774 -0.175     0.0204 smoothed in_sample
#> 5 2003-10-01 news_vint1  -0.153  -0.251    -0.0555 smoothed in_sample
#> 6 2004-01-01 news_vint1  -0.185  -0.283    -0.0876 smoothed in_sample
head(states(fit, filter = "filtered", state = "true_lag_0"))
#> # A tibble: 6 × 7
#>   time       state      estimate   lower   upper filter   sample   
#>   <date>     <chr>         <dbl>   <dbl>   <dbl> <chr>    <chr>    
#> 1 2002-10-01 true_lag_0  0.0577  -0.0403 0.156   filtered in_sample
#> 2 2003-01-01 true_lag_0 -0.00601 -0.104  0.0920  filtered in_sample
#> 3 2003-04-01 true_lag_0 -0.0957  -0.194  0.00226 filtered in_sample
#> 4 2003-07-01 true_lag_0  0.461    0.363  0.559   filtered in_sample
#> 5 2003-10-01 true_lag_0  0.465    0.367  0.563   filtered in_sample
#> 6 2004-01-01 true_lag_0  0.758    0.660  0.856   filtered in_sample
# }