Hot deck to impute wind series

[jorge quishpe]

hello with everyone
I am a new user in SATA, I investigated a bit about the imputation of data and I saw that the HOT DECK method was very useful, I saw that there is a library or a command to be able to impute the data, but I am not sure how is it done or how does it work in if the codifo, if anyone knows please could you assist me with your help thank you

[Carlo Lazzaro]

Jorge:
welcome to this forum.
The hot deck approach (ie, drawing a random value from the ones observed and use it to replace a missing value in the same dataset) is a bit old fashioned and suffers from some drawbacks, the most relevant being the absence of variance of the imputed value.
A preferable approach is multiple imputation, that Stata offers under -mi- entries (see the Stata .pdf manual).

[jorge quishpe]

Hi Carlos, I think I will have a problem since I must use the hot deck method for an article, I read several articles where I found that the Hot Deck method also makes a multiple imputation, you know something about it. thanks for your help

[Carlo Lazzaro]

Jorge:
you can find an interesting description of the single imputation method called "hot deck" at https://www.guilford.com/books/Missi.../9781593853938, pages 182-184.

[jorge quishpe]

Carlo: I could not see the book you sent me, since it has a cost, if you could help me with this method or how to use it in STATA it would be very helpful,

[Carlo Lazzaro]

Jorge:
see: https://www.ncbi.nlm.nih.gov/pmc/art...ihms299474.pdf.
Sorry I cannot be more helpful.

[skolenik]

jorge quishpe , you need to explain how your data are set up. I suspect that wind data are highly skewed and irregular, and hot deck is a method that only works with highly stable and easy to predict data that can be fully explained by less than a handful of categorical variables.