A New Framework of Feature Engineering for Machine Learning in Financial Fraud Detection

Abstract

Financial fraud activities have soared despite the advancement of fraud detection models empowered by machine learning (ML). To address this issue, we propose a new framework of feature engineering for ML models. The framework consists of feature creation that combines feature aggregation and feature transformation, and feature selection that accommodates a variety of ML algorithms. To illustrate the effectiveness of the framework, we conduct an experiment using an actual financial transaction dataset and show that the framework significantly improves the performance of ML fraud detection models. Specifically, all the ML models complemented by a feature set generated from our framework surpass the same models without such a feature set by nearly 40% on the F1-measure and 20% on the Area Under the Curve (AUC) value.

New Feature Engineering Framework for Deep Learning in Financial Fraud Detection

“…In the new framework [19], we use feature aggregation and feature transformation jointly to create important feature candidates. R. Wedge et al [15] suggest Deep Feature Synthesis (DFS) that creates new attributes for machine learning models of credit card fraud detection using the relational structure of the dataset.…”

“…In the work described in [19], we used the engineered features created through using the feature engineering framework which has improved the performance of machine learning models.…”

“…From our previous published work in [19], it has been demonstrated that the performance of fraud detection models with prepared feature sets, performs better than the models with original data only.…”

“…In the new feature engineering framework published in [19], we created and selected the effect features for fraud detection models built with ML algorithms: We selected Support Vector Machine (SVM) and Isolation Forest (IF) as fraud detection models.…”