Timeline for Forecasting using functional data analysis with parametric models for the functions
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 5, 2022 at 21:59 | comment | added | Rob Hyndman | The first paper on this was Hyndman & Ullah (CSDA, 2007). Citations of that paper, including the one Stephan mentions, may also be useful. | |
| Jul 5, 2022 at 18:28 | comment | added | Aksakal | what you describes sounds like a variation of state space modeling. you essentially posit that the observed data is a representation of some unobserved state variables. so your space of x's, observed, are a transformation with noise x=f(s)+e of unobserved state s. you try to estimate s, then forecast $\hat s$ and transform back to $\hat x$. there are many examples of this approach in different fields. | |
| Jul 5, 2022 at 18:14 | history | edited | Barry | CC BY-SA 4.0 |
deleted 2 characters in body
|
| Jul 5, 2022 at 17:30 | comment | added | Stephan Kolassa | Forecasting of density function is not a duplicate (it asks for examples, not algorithms), but it might be useful. Unfortunately, there is not a lot of academic work on functional data forecasting. The Hyndman & Shang (2009) paper I referred to at this other thread is the only such paper I am aware of. Searching the agenda of the upcoming International Symposium on Forecasting for "functional" yields a few hits. Consider contacting the presenters? | |
| Jul 5, 2022 at 17:22 | history | asked | Barry | CC BY-SA 4.0 |