DBNex: Deep Belief Network and Explainable AI based Financial Fraud Detection

Abhimanyu Bhowmik, Madhushree Sannigrahi, Deepraj Chowdhury, Ashutosh Dhar Dwivedi, Raghava Rao Mukkamala

Abstract: The majority of financial transactions are now conducted virtually around the world. The widespread use of credit cards and online transactions encourages fraudulent activity. Thus, one of the most demanding real-world challenges is fraud detection. Unbalanced datasets, in which there are a disproportionately high number of non-fraud samples compared to incidents of fraud, are one of the key obstacles to effective fraud detection. A further factor complicating the learning process for cutting-edge machine learning classifiers is how quickly fraud behaviour changes. Thus, in this study, we suggest an efficient fraud detection methodology. We propose a unique nonlinear embedded clustering to resolve imbalances in the dataset, followed by a Deep Belief Network for detecting fraudulent transactions.The proposed model achieved an accuracy of 94 % with a 70:30 ratio of training-validation dataset.

About Dataset

Context

The Credit Card Fraud Detection dataset by Worldline and the Machine Learning Group of ULB (Universit´e Libre de Bruxelles) is used in this paper. It includes labelled anonymized credit card payments made by European card users in September 2013.

Content

The dataset contains 492 frauds out of 284,807 transactions that took place over the course of 2 days.

Dataset Specifications:

• PCA (Principal Component Analysis) transformed into features V1, V2,... V28
• Time
• Amount
• Class

Note: Only 0.172% of the transactions in the dataset are fraudulent,
which indicates a severe lack of balance.

Sample Dataset from the actual dataset can be found in the the folder ./Data_and_Model.

Proposed Methods

1. Globar Architecture

Overall view of the proposed model: DBNex

2. Data Preprocessing

Uniform Manifold Approximation and Projection (UMAP)

Clustering of UMAP-based 2D embedding of non-fraud samples using DBSCAN

Density-Based Spatial Clustering of Applications with Noise or DBSCAN

Silhouette score of DBSCAN Clusters for non-fraud samples where ϵ ranging from 0.1 to 0.6.

3. Proposed Model

Architecture of Deep Belief Network

Results

1. Classification Report

Classification report for 70:30 train-validation ratio

2. Confusion Matrix

Confusion matrix: Fraud-NonFraud classes

3. ROC - AUC Curve

Receiver operating characteristic curves: Fraud-NonFraud classes

4. Classification Explainability using SHAP

(a)

(b)

(a), (b) Beeswarm plot of testing dataset for both the classes (left) Heatmap plot of testing dataset for both the classes (right).

Cite our work

  @INPROCEEDINGS{10020494,
  author={Bhowmik, Abhimanyu and Sannigrahi, Madhushree and Chowdhury, Deepraj and Dwivedi, Ashutosh Dhar and Rao Mukkamala, Raghava},
  booktitle={2022 IEEE International Conference on Big Data (Big Data)}, 
  title={DBNex: Deep Belief Network and Explainable AI based Financial Fraud Detection}, 
  year={2022},
  volume={},
  number={},
  pages={3033-3042},
  doi={10.1109/BigData55660.2022.10020494}}

Contact