A gradient boosting example with
xgboost library to reveal suspicious transactions. We will be working with a transaction log dataset from Synthetic Financial Datasets For Fraud Detection.
For experimentation tracking, we will use the area under the precision-recall curve (AUPRC) rather than the conventional area under the receiver operating characteristic (AUROC), since the data is highly skewed.
- Feature Store: We are going to use SQL queries to build the
transactionfeatures and use it to train the
- Experimentation tracking
- logging parameters:
To check out the Layer Fraud Detection example, run:
To build the project: