Basic functionality • pickmdl3

In this vignette we explore the basic function x13_pickmdl. As its name indicates, this function is a wrapper function that runs the x13 function, but with pickdml rather than automodel as the default model selection procedure.

We first load a time series to be seasonally adjusted. The series in question is the Norwegian retail index for nace 47.6, , between January 2014 and July 2026. We define a specification to be used in the example, namely the default specification ‘rsa5c’, where we turn off transtive outliers.

Just as in the x13 function, x13_pickmdl takes the time series and the specification as inputs. Output is a list with fitted model results, where the most important information is shown in the console. We see that the model of order $(2,1,0)(0,1,1)_s$ is selected.


library(rjd3toolkit)
library(rjd3x13)
library(pickmdl3)

rti_476 <- pickmdl3::pickmdl_data("norwegian_rti")$rti_476

spec_now <- x13_spec("rsa5c") |>
  set_outlier(outliers.type=c("AO","LS")) 

model_476 <- pickmdl3::x13_pickmdl(ts = rti_476,spec=spec_now)

model_476
#> Serie span: All 
#> 
#> Model: X-13
#> Log-transformation: yes 
#> SARIMA model: (2,1,0) (0,1,1)
#> 
#> SARIMA coefficients:
#>    phi(1)    phi(2) btheta(1) 
#>    0.5141    0.1056   -0.4443 
#> 
#> Regression model:
#>             mon             tue             wed             thu             fri 
#>      -0.0107469       0.0062918       0.0099814      -0.0068710       0.0190396 
#>             sat AO (2020-01-01) AO (2020-03-01) LS (2020-05-01) LS (2020-08-01) 
#>       0.0007883      -0.1749130      -0.2249022       0.2078185      -0.1806520 
#> LS (2021-05-01) LS (2021-08-01) 
#>       0.2242468      -0.1430262 
#> 
#>  Seasonal filter: FILTER_S3X5;  Trend filter: H-13 terms
#>  M-Statistics: q Good (0.393); q-m2 Good (0.433)
#>  QS test on SA: Good (0.975);  F-test on SA: Good (0.547)
#> 
#> For a more detailed output, use the 'summary()' function.

Further information about the model choiche can be retrieved with the function ok, which gives the user a list with three objects. The first object $ok tells the user whether the selected model fulfills the three pickmdl criteria at the moment of model selection. The second object ok_final says wheter the selected model fulfills the three pickmdl criteria at the end of the time series. The third object mdl_nr says which model on the pickmdl list that was selected. We see that the third model was selected, which of course is the model of order $(2,1,0)(0,1,1)_s$ .

pickmdl3::ok(model_476)  
#> $ok
#> [1] TRUE
#> 
#> $ok_final
#> [1] TRUE
#> 
#> $mdl_nr
#> [1] 3

The calculated pickmodel criteria are included in the output if the option output is set to “all”. Now we can see that the first two models were rejected because they didn’t pass the second criterium, which is indepedent residuals.

model_476 <- pickmdl3::x13_pickmdl(ts = rti_476,spec=spec_now,output = "all")  
model_476$crit_tab
#>           crit1      crit2      crit3    m_aic
#> [1,] 0.03655921 0.02228086 -0.7916005 860.9283
#> [2,] 0.03546483 0.02464104 -0.7901148 831.0459
#> [3,] 0.03107798 0.64725040  0.0000000 788.2761

Note that there are only three models on the list. This is because the algorithm by default stops at the first model that passes the criteria. To force calculation of all five models, set the option fastfirst = FALSE:

model_476 <- pickmdl3::x13_pickmdl(ts = rti_476,spec=spec_now,output = "all",
                                   fastfirst = FALSE)
model_476$crit_tab
#>           crit1       crit2      crit3    m_aic
#> [1,] 0.03655921 0.022280863 -0.7916005 860.9283
#> [2,] 0.03546483 0.024641045 -0.7901148 831.0459
#> [3,] 0.03107798 0.647250400  0.0000000 788.2761
#> [4,] 0.03154734 0.177456254 -0.9999757 852.7246
#> [5,] 0.03483096 0.008379515 -0.7677882 838.9111

When the option output = "all", the selected model is available in $sa. A list of all the calculated models in the pickmdl list are given in $all.

model_476$sa
model_476$sa_mult

Finally, the automodel procedure maybe used also with the x13_pickmdl function. When automdl.enabled = TRUE the automodel procedure is selected, which gives the exact same result as the x13 function.

model_476_auto <- pickmdl3::x13_pickmdl(ts = rti_476, spec = spec_now, automdl.enabled = TRUE)
model_476_auto 
#> Serie span: All 
#> 
#> Model: X-13
#> Log-transformation: yes 
#> SARIMA model: (1,0,1) (0,1,1)
#> 
#> SARIMA coefficients:
#>    phi(1)  theta(1) btheta(1) 
#>   -0.7505   -0.2654   -0.3421 
#> 
#> Regression model:
#>             mon             tue             wed             thu             fri 
#>       -0.012691        0.012751        0.005173       -0.011958        0.023985 
#>             sat LS (2020-02-01) AO (2020-03-01) 
#>        0.001713        0.164992       -0.270125 
#> 
#>  Seasonal filter: FILTER_S3X3;  Trend filter: H-13 terms
#>  M-Statistics: q Good (0.502); q-m2 Good (0.559)
#>  QS test on SA: Good (1.000);  F-test on SA: Good (0.687)
#> 
#> For a more detailed output, use the 'summary()' function.