Repository for reproducing the results presented in the paper "Tuning gradient boosting for imbalanced bioassay modelling".
- losses: folder containing the implementation of the custom loss functions and related classes.
- misc: folder containing utility functions to run the benchmarks.
- scripts: collection of scripts to reproduce the results presented in the paper.
- output: placeholder to store the outputs of the pipelines in scripts.
- env.yml: conda environment with all packages required to run the scripts.
- datasets.zip .csv files of all publicly available datasets evaluated in the paper, has to be extracted to run the scripts.
New version of the package available at https://github.com/dahvida/custom_gbm
Please reference the original publication: https://jcheminf.biomedcentral.com/articles/10.1186/s13321-022-00657-w