Overview

Problem

The acceleration of digital transformation creates an urgent need to use electronic payment, and thus fraudsters use highly technical methods

Objectives

• Implement a real time fraud detecting model.
• Improve the accuracy of the most recent findings.
• Raise the awareness of the fraudulent activities associated with the online payments process.

Solution Used

In our system, detecting avoids any attacks from fraudsters by acting as a layer of protection. Any transaction that is put under the fraud category is rejected. Otherwise, the transaction gets approved.

Algorithms are GBC, XGB, RF, on addition to the ANN due to the Concept Drift, which indicates the change in the data over time.

Methodology

Workflow

• Data Collection
• Preprocessing
• Analyzing
• Model Formulation and Validation
• Run and Test

FDS App

During annual exhibition of IT college at University of Babylon, I showed up this work in a website executed by https://www.gradio.app/.

https://github.com/sajjadhasann/Fraud-Detection-Sytem

https://youtu.be/gwnYSVEp8mM