-
-
Notifications
You must be signed in to change notification settings - Fork 5
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
add recent power values as features #115
Conversation
elif len(recent_power_values) > self._n_recent_power_values: | ||
recent_power_values = recent_power_values[ | ||
len(recent_power_values) - self._n_recent_power_values : | ||
] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Changes might need to be made when we use the flag --no-live-pv, to remove PV features during inference for metadata only backtests.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yea, good shout, ill have a look at that. Probably the same for the mean power values
above.
Although the default is 0 values, so it might be ok
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The new features would need to be removed from here:
pv-site-prediction/psp/models/multi.py
Line 30 in 7a46fd0
def get_features_without_pv(self, x: X, is_training: bool = False) -> Features: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
currently these pv features are hardcoded (in line 48):
pv_derived_features = ["recent_power", "h_max", "h_median", "h_mean"]
nan_pv_derived_features = ["recent_power_nan", "h_max_nan", "h_median_nan", "h_mean_nan"]
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think I found it
Looks good. It would probably make sense to add a test somewhere for the --no-live-pv. (which is my bad as I didn't make one when i first implemented this!) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good
Pull Request
Description
Add
recent values values
as a feature. Before we jsut have the recent power mean is a recent time period.TODO run experiment with and without this. The experiment should perhaps just be trying to forecast 1 hour in 15 mins blocks. And we would expect this to make a difference
How Has This Been Tested?
CI tests
Checklist: