Unsigned
7K+
4
fraud-detection
:Use the Pull Tag button to download this ModelKit.
Or, read our KitOps documentation to learn how to use kit unpack --filter to download only the components you need.
ModelKit Tag Metadata
- Author
- Jozu
- Date added
- Size
- 24.9MB
- Digest
- Total pulls
- 7
Package
- Name
- fraud-detector
- Version
- 2.1
- Authors
- Jozu
- Description
- An ML model to identify fraudulent transactions
Model
- Name
- fraud-xgb
- Path
- model/model.pkl
- License
- Apache 2.0
- Parts
- N/A
- Parameters
{
"MLFlow": {
"run_info": {
"run_id": "611dd98a4fbc42f78c20c559337f8202",
"run_name": "industrious-loon-433",
"artifact_uri": "mlflow-artifacts:/693112939931118568/611dd98a4fbc42f78c20c559337f8202/artifacts",
"experiment_id": "693112939931118568"
},
"result_info": {
"metrics": {
"score": 0.649,
"roc_auc": 0.83,
"f1_score": 0.64,
"log_loss": 0.902,
"recall_score": 0.649,
"example_count": 1617,
"accuracy_score": 0.649,
"precision_score": 0.652
}
},
"best_hyperparameters": {
"max_depth": 10,
"n_estimators": 50,
"learning_rate": 0.1
}
}
}