feat: models #3
feat: models #3
Conversation
corymosiman12
left a comment
corymosiman12
left a comment
There was a problem hiding this comment.
Minor changes - seems simple enough.
| "Daily" = "Day", | ||
| "Hourly" = "Hour", | ||
| "SLR" = "2 parameter simple linear regression", | ||
| "Three Parameter Heating" = "3 parameter heating change point model", |
There was a problem hiding this comment.
@macintoshpie should we remove the redundant ones? i.e. I assume the keys with 'Three Parameter Heating' and '3PH' are the same, and you are probably only using '3PH'...unless I missed something.
There was a problem hiding this comment.
We need both from what I can tell b/c it depends on what you pass to nmecr (which allows both)
model <- nmecr::model_with_CP(bsyncr_df,
nmecr::assign_model_inputs(regression_type = "Three Parameter Cooling"))
print(model$model_input_options$regression_type)
# prints "Three Parameter Cooling"
model <- nmecr::model_with_CP(bsyncr_df,
nmecr::assign_model_inputs(regression_type = "3PC"))
print(model$model_input_options$regression_type)
# prints "3PC"
corymosiman12
left a comment
corymosiman12
left a comment
There was a problem hiding this comment.
Looks good. Anyway to add some testing? Or just add an issue to begin define testing? I think we can probably just have a predefined:
- Input file
- Output file (expected)
And then the test just compares a new output file to the expected output file. It can be more 'outcome based' than unit test based - make sense?
|
yeah we should definitely add some testing at some point |
Adds ability to export 3 and 4 parameter models into bsync.
5P is not handled b/c it is failing to complete with our SEED test data currently.