Mainak Jas

flake8 on tests

2

The flake8 linter doesn't seem to be working on the test directory. @pavanramkumar did we disable it at some point? In any case, we should fix this as soon as...

Implement sample weight

1

@peterfoley do you mind if we borrow code from your fork for `pyglmnet`? :-) Probably won't get merged soon but I wanted to open a PR so that it's on...

[MRG] FIX problem of learning rate

9

closes https://github.com/glm-tools/pyglmnet/issues/65 @pavanramkumar I am not sure if I got the lipschitz constant for logistic regression right, and we need to dig up the other Lipschitz constants or see how...

In most of the other projects I contribute to, we use `pydocstyle` to check conformity to PEP257. Once #259 is in, we can enable it by simply switching the line...

[MRG] Improve test coverage

8

Let's see if this helps with coverage ...

translate coordinate descent to cython

12

I just chatted with @rvraghav93 and he is willing to help us with cython code. This might be a good starting point: https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/linear_model/cd_fast.pyx it might also be easier to benchmark...

closes #211 @pavanramkumar this one is not ready though. Tests do not pass ...

We should refactor the hessian out of the class and have test the same way we have for the gradients

we should use vectorized operations., e.g., `np.einsum` or `np.dot` instead of `np.sum(x * y)`

validation of labels

1

We should validate the labels y during `fit`. For example, if it's `binomial` allow only 0 and 1 or only integers if it's `poisson` etc ...