fairgbm
fairgbm copied to clipboard
feedzai
Train Gradient Boosting models that are both high-performance *and* Fair!
- LightGBM makes available their official documentation at https://lightgbm.readthedocs.io/en/v3.3.2/ - FairGBM features changes to the C API, new parameters, and other changes to the Python API - We should build...
Closes #6 **TODO**: as per the profiling we ran, the first thing to optimize is all the string comparisons happening in the CosntrainedObjectiveFunction -- these take up around 1/3rd of...
## Description - When optimizing for Recall (minimizing FNR), only label positive samples are considered for computing the loss or its gradient; - However, passing a gradient of zero for...
## Summary Currently, the function `ConstrainedObjectiveFunction::GetLagrangianGradientsWRTMultipliers` ignored the `this->weights_` variable. **TODO**: implement weighing -- although this may interfere with the constrained optimization process.
## Summary LightGBM allegedly handles missing values in the features (e.g., represented as `NaN`). We should also be able to handle missing values in the constraint group column (sensitive attribute)....
As people noticed (https://github.com/feedzai/fairgbm/issues/45), fairgbm is not supported on macOS. Besides, I tested fairgbm on Cent OS and RHEL8 systems, and it didn't work as well. So I don't think...
my task is regression，but I cannot find the python API FairGBMRegressor. Is there a plan to develop it?
