DingLuoXMU / XGraphBoost

XGNNBOOST

XGBOOST is an algorithm combining GNN and XGBOOST, which can introduce the machine learning algorithm XGBOOST under the existing GNN network architecture to improve the algorithm capability.The GNN used in this paper includes DMPNN, GGNN and GCN

Installation

Create a conda environment :

conda create env -f environment.yaml

conda activate my-rdkit-env

train--dmpnn_xgboost

cd /home/xgraphboost/dmpnn_xgboost
python train.py --protein delaney_ESOL --dataset_type regression --save_dir delaney_ESOL_r --epoch 200

train--gcn_xgboost

cd /home/xgraphboost/gcn_ggnn_xgboost
python gcn_xgb_morgan.py --folder benchmark_molnet/molnet_random_bbbp_c --type classification

train--ggnn_xgboost

cd /home/xgraphboost/gcn_ggnn_xgboost
python xgb_ggnn.py --folder benchmark_molnet/molnet_random_bbbp_c --type classification