You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Jun 2, 2023. It is now read-only.
We had been training models with some GAGESII dataset attributes. That will not allow us to make predictions for all NHD reaches because those attributes will not be available.
There are similar NHD attributes available for most of these:
Freshwater withdrawal - is this available?
NPDES - we can pull this additional dataset.
development (devl) is already captured from the SOHL data (highly correlated)
Canal and ditch data are already pulled, but seemed better from GAGESII compared to the NHD data (fewer 0s in the record). We can switch back to the NHD dataset for prediction in all reaches.
For now, I'll re-run our models without these GAGESII data. I'm not expecting any differences for our multiclass prediction model because these attributes were not in the top most important.
The text was updated successfully, but these errors were encountered:
We had been training models with some GAGESII dataset attributes. That will not allow us to make predictions for all NHD reaches because those attributes will not be available.
There are similar NHD attributes available for most of these:
For now, I'll re-run our models without these GAGESII data. I'm not expecting any differences for our multiclass prediction model because these attributes were not in the top most important.
The text was updated successfully, but these errors were encountered: