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
SKLL pull request #110 adds support to SKLL for printing SVR model weights and corrects the sign. However, whenever sklearn is fixed, the correction should be removed.
The text was updated successfully, but these errors were encountered:
It looks like they've fixed scikit-learn/scikit-learn#2933, but it's not released yet. We should keep in mind that whenever the next release of scikit-learn comes out, we'll have to fix this.
- Remove SVR coefficient correction (#111) now that the original bug has been fixed in sklearn.
- Remove dependency on `BaseLibLinear` in `model_params()` since it is no longer exposed in scikit-learn.
- Expose `LinearSVR` from scikit-learn and so do not have SVR have a 'linear' kernel by default.
- Include `RescaledLinearSVR` among the rescaled regressors.
- Remove `SVR` from `test_linear_models()` in `test_regression.py` and create a new `test_non_linear_models()`.
- Remove SVR coefficient correction (#111) now that the original bug has been fixed in sklearn.
- Remove dependency on `BaseLibLinear` in `model_params()` since it is no longer exposed in scikit-learn.
- Expose `LinearSVR` from scikit-learn and so do not have SVR have a 'linear' kernel by default.
- Include `RescaledLinearSVR` among the rescaled regressors.
- Remove `SVR` from `test_linear_models()` in `test_regression.py` and create a new `test_non_linear_models()`.
Currently, sklearn appears to output coefficients with the wrong sign for SVR. See scikit-learn/scikit-learn#2933.
SKLL pull request #110 adds support to SKLL for printing SVR model weights and corrects the sign. However, whenever sklearn is fixed, the correction should be removed.
The text was updated successfully, but these errors were encountered: