-
Notifications
You must be signed in to change notification settings - Fork 1
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
Test accuracy of predict
methods
#26
Comments
@aazuspan, a quick note. I figured out my issue with the predictions and it was how I was calling yaImpute. I've got some tests for both equal-weights and distance-weighted predictions and they are all passing. Woot! Note that when you call |
That's great news @grovduck! Agreed that Thanks for the code reference and the explanation of the issue! |
I'd say go ahead as is! |
We exported data from yaImpute for the five core estimators at k=5 for both unweighted (equal-weighting to all neighbors) and weighted (neighbors are weighted by their inverse distance) predictions of the y attributes. The tests assert that the predict method in each estimator is returning the same values. Note that the constructor of each estimator uses a custom function (yaimpute_weights) to the weights parameter in order to match the weighting scheme of yaImpute.
We exported data from yaImpute for the five core estimators at k=5 for both unweighted (equal-weighting to all neighbors) and weighted (neighbors are weighted by their inverse distance) predictions of the y attributes. The tests assert that the predict method in each estimator is returning the same values. Note that the constructor of each estimator uses a custom function (yaimpute_weights) to the weights parameter in order to match the weighting scheme of yaImpute. This commit fixes #26.
We exported data from yaImpute for the five core estimators at k=5 for both unweighted (equal-weighting to all neighbors) and weighted (neighbors are weighted by their inverse distance) predictions of the y attributes. The tests assert that the predict method in each estimator is returning the same values. Note that the constructor of each estimator uses a custom function (yaimpute_weights) to the weights parameter in order to match the weighting scheme of yaImpute. This commit fixes #26.
Resolved by #32 |
Currently, we test the accuracy of
fit
andkneighbors
againstyaImpute
in test_port.py, but we don't yet test the accuracy ofpredict
, which makes it easy for small mistakes that affect results but not functionality (like #23) to go unnoticed.As discussed in that issue, we should add some tests that compare prediction results of our estimators to those generated with
yaImpute
.The text was updated successfully, but these errors were encountered: