openclimatefix / power_perceiver

Machine learning experiments using the Perceiver IO model to forecast the electricity system (starting with solar)
MIT License
7 stars 1 forks source link

[META] forecast 8 hrs into the future #133

Closed JackKelly closed 2 years ago

JackKelly commented 2 years ago
JackKelly commented 2 years ago

Maybe the simplest thing to do is to not filter by the Sun's angle at all, and maybe extend the forecast horizon to, like, 24 hours so every example is guaranteed to have some daylight and nighttime. Although, of course, the HRV imagery won't show anything useful at night!

So each example could start at literally any time of day (00:00, 06:55, 12:00, etc.)

JackKelly commented 2 years ago

Any thoughts about the idea in the comment immediately above this comment, @peterdudfield and @jacobbieker?

peterdudfield commented 2 years ago

I think its not crazy, to just do full 24 hours start time. Could always look into it if it really starts effecting performance? But as a good default, probably worth trying

peterdudfield commented 2 years ago

The your MAE will probably go down? easier to predict at night. Have to think how we compare that properly

peterdudfield commented 2 years ago

I would stick to a 8 hour horizon at the moment though, just 24 hours start time.

let's get that working first, and then we can think about extending

JackKelly commented 2 years ago

All sounds good!

And, yeah, I'd need to make sure the validation metrics are computed in the same way (using only daylight hours).

jacobbieker commented 2 years ago

Sorry for being a bit delayed with this, but yeah, agree with peter, just have it start at any time and the 8 hour horizon,