Functions — github.com/RussanaMary/Fraud-Detection

CodeHub MCPcopy Create free account

hub / github.com/RussanaMary/Fraud-Detection / functions

Functions10 in github.com/RussanaMary/Fraud-Detection

⨍ Functions 10 ◇ Types & classes 0

↓ 12 callersFunctioncard_ax

↓ 1 callersFunctionevaluate_models

Computes per-model: ROC-AUC, PR-AUC, F1, Precision, Recall, Confusion Matrix, and full classification report.

↓ 1 callersFunctiongenerate_report

(results: dict, df: pd.DataFrame)

↓ 1 callersFunctiongenerate_transaction_data

Generates a realistic synthetic transaction dataset with: - 10 PCA-style anonymised features (V1-V10) - 9 domain-engineered fraud indi

↓ 1 callersFunctionmain

↓ 1 callersFunctionmake_dashboard

(df: pd.DataFrame, results: dict, feature_cols: list, y_test: pd.Series)

↓ 1 callersFunctionmanual_oversample

Interpolates synthetic minority-class samples between existing fraud rows until the minority class is `ratio` × size of majority class.

↓ 1 callersFunctionplot_roc_pr

(results: dict, y_test: pd.Series)

↓ 1 callersFunctionpreprocess

- Label-encode categorical column - StandardScale continuous features - Stratified 80/20 train-test split

↓ 1 callersFunctiontrain_models

Trains four classifiers, each with class_weight='balanced' to further penalise misclassification of the minority class.