Summary
The Dask estimators in lightgbm.dask should support the use of a custom objective function.
Motivation
This feature would bring Dask estimators closer to parity with the sklearn estimators.
Description
I haven't thought through this much yet, just writing up a placeholder issue for discussion. If you're reading this and have ideas, please comment and the issue can be re-opened.
References
See
|
A custom objective function can be provided for the ``objective`` parameter. |
|
In this case, it should have the signature |
|
``objective(y_true, y_pred) -> grad, hess`` or |
|
``objective(y_true, y_pred, group) -> grad, hess``: |
|
|
|
y_true : array-like of shape = [n_samples] |
|
The target values. |
|
y_pred : array-like of shape = [n_samples] or shape = [n_samples * n_classes] (for multi-class task) |
|
The predicted values. |
|
group : array-like |
|
Group/query data. |
|
Only used in the learning-to-rank task. |
|
sum(group) = n_samples. |
|
For example, if you have a 100-document dataset with ``group = [10, 20, 40, 10, 10, 10]``, that means that you have 6 groups, |
|
where the first 10 records are in the first group, records 11-30 are in the second group, records 31-70 are in the third group, etc. |
|
grad : array-like of shape = [n_samples] or shape = [n_samples * n_classes] (for multi-class task) |
|
The value of the first order derivative (gradient) for each sample point. |
|
hess : array-like of shape = [n_samples] or shape = [n_samples * n_classes] (for multi-class task) |
|
The value of the second order derivative (Hessian) for each sample point. |
|
|
|
For binary task, the y_pred is margin. |
|
For multi-class task, the y_pred is group by class_id first, then group by row_id. |
|
If you want to get i-th row y_pred in j-th class, the access way is y_pred[j * num_data + i] |
|
and you should group grad and hess in this way as well. |
for an explanation of how this works in the sklearn estimators.
Can look at how xgboost.dask handles this for inspiration:
