Scribd
Upload
13 views2 pages

AI-Driven E-Commerce Fraud Detection

The document discusses a novel hybrid approach to combat fraudulent transactions in e-commerce by integrating process mining and machine learning for real-time user behavior analysis. It highlights the limitations of existing systems that rely on historical data and proposes a system that utilizes SVMs and multi-perspective analysis for enhanced fraud detection. Key features include dynamic monitoring, anomaly detection, and improved visualization of transaction flows.

Uploaded by

manojachitturi5783
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
13 views2 pages

AI-Driven E-Commerce Fraud Detection

The document discusses a novel hybrid approach to combat fraudulent transactions in e-commerce by integrating process mining and machine learning for real-time user behavior analysis. It highlights the limitations of existing systems that rely on historical data and proposes a system that utilizes SVMs and multi-perspective analysis for enhanced fraud detection. Key features include dynamic monitoring, anomaly detection, and improved visualization of transaction flows.

Uploaded by

manojachitturi5783
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Using Artificial Intelligence to Combat COVID-19

Problem Statement/Abstract

The document highlights the challenge of detecting and preventing fraudulent transactions in

e-commerce platforms. Current systems struggle due to their reliance on historical data and lack of

dynamic process monitoring. To address this, the study proposes a novel hybrid approach that

integrates process mining and machine learning to analyze user behaviors in real-time. Key

components include process modeling, anomaly analysis, and a Support Vector Machine

(SVM)-based classification model to effectively detect fraudulent transactions.

Technical Requirements

1. Conformance Checking: Implemented using process mining tools like Petri nets to detect anomalies.

2. Machine Learning: Utilizes SVMs to classify fraudulent behaviors based on multi-perspective data.

3. Event Logs: Essential for analyzing transaction flows and extracting behavioral features.

4. Software Tools: Tools such as ProM, Disco, and Heuristic Miner are mentioned for data analysis and con

Existing System

1. Primarily relies on historical data for detecting fraud.

2. Utilizes supervised machine learning for identifying abnormal user behavior.

3. Focuses on credit card fraud detection but has limited adaptability to dynamically changing

fraudulent behaviors.

4. Process mining has been used but lacks integration with real-time multi-perspective analysis.

Disadvantages:

- Fraudulent activities, such as tampered orders and subcontracted payments, evade detection.

- Existing methods are static and fail to capture real-time behavioral changes in e-commerce transactions.

Proposed System

The new system combines process mining and machine learning to enhance fraud detection:
1. Real-Time Behavior Analysis: Monitors and analyzes user behaviors dynamically.

2. Multi-Perspective Analysis: Examines transactions across control flow, resource, time, and data

perspectives.

3. Fraud Detection Model: SVM embedded with multi-perspective process mining.

4. Enhanced Visualization: Uses color-coded data flow models to identify mismatches and

anomalies.

Advantages:

- Provides a clearer result with tools like Multi-Perspective Process Explorer.

- Efficiently identifies mismatches in transactional data for improved fraud detection.

You might also like