GradientBoostingClassifier and GradientBoostingRegressor should support sparse matrices

[SimengSun]

MemoryError will be raised for large training data when using GBC (It converts sparse training data to dense first). GBC in sklearn does support sparse matrices as input and works well.

[desilinguist]

SKLL currently expects the following learners to take only dense matrices as input:

_REQUIRES_DENSE = (BayesianRidge,
                   GradientBoostingClassifier,
                   GradientBoostingRegressor,
                   Lars,
                   TheilSenRegressor)

I double checked and other than the gradient boosted learners, the others take dense matrices as expected.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.

[desilinguist]

Keep it open please.

[desilinguist]

Addressed by #429.