Fraud Detection with EvalML
Jan 1, 2022
Data analytics has created a great impact in the banking and financial services industry, for example, by providing insights of global financial trends and financial modelling etc. Among them, fraud prevention and detection are one of the applications. This article applied predictive data analytics and supervised machine learning (ML) methods for card-not-present (CNP) fraud detection, and demonstrated modelling using EvalML , an auto machine learning library. This article also identified that both Decision Tree (DT) and XGBoost models work better than Linear models (LM), Random Forest (RF) and LightGBM models.
The dataset used to demonstrate modelling is a large-scale dataset from Vesta which is available on Kaggle .