Does someone know the function to create density forecasts within the GAMLSS Package?
The predict. Formula is not the right one. Predict do Point Forecasts
1 Answer
$\begingroup$
$\endgroup$
3
GAMLSS uses distributional assumptions, and estimates and forecasts conditional distribution parameters. So you would need to use the what parameter to predict.gamlss to tell it which distribution parameters to predict, e.g., mu, sigma, nu or tau. (Actually, you will probably want to use type=response, because the default type=link gives you the logit prediction.) Then feed these parameters into the distribution you used originally in fitting the GAMLSS model, and use the appropriate density function.
answered Jan 29 at 12:28
-
$\begingroup$ Thank you for answer. How do I feed the parameters into the distribution? And how do I find the appropriate density function? $\endgroup$– NushkoCommented Jan 30 at 13:05
-
$\begingroup$ That will depend on what you fitted in the first place per
gamlss.famliy. If you could edit your question to include a Minimum Working Example (usedput()!) and notify me here in another comment, I will try to take a look. $\endgroup$ Commented Jan 30 at 13:10 -
$\begingroup$ Sorry for the late answer, but I found the right function. Its the prodist() Function $\endgroup$– NushkoCommented Feb 2 at 10:42