Gamma regression does not work with example in docs

[ThomasMGeo]

X, Y, T = xc.load_sample_data() # load test data

gamma = xc.rGammaRegression()
gamma.fit(X, Y)
pred = gamma.predict(X)

ValueError: Some value(s) of y are out of the valid range of the loss 'HalfGammaLoss'.

[kjhall01]

Hi Thomas,

https://github.com/scikit-learn/scikit-learn/blob/c900ad385cecf0063ddd2d78883b0ea0c99cd835/sklearn/linear_model/_glm/glm.py#L116

Scikit-Learn's Gamma Regression "HalfGammaLoss" loss function requires data to be greater than 0. there are some gridpoints with precipitation = 0 in the sample data, so It shouldn't work.

Some options might be:

1/ implement a wrapper for the 2D gamma regressor implementation which drops 0-valued samples , then use XCast to extend it
2/ add something very small to Y before fitting - like: gamma.fit(X, Y+0.0001)

You may want to do option 1 anyway, because it looks like gamma regressor needs a different shape input vector than normal. Happy to help if you'd like.

[ThomasMGeo]

Option 2 does work, agree that 1 is a better path. I will work on it for a bit, and reach out if I get stuck.

Thanks!