This document proposes a mechanism to detect credit card fraud in online transactions using a Hidden Markov Model. The model would classify users as having low, medium, or high spending habits and flag transactions as potentially fraudulent if a user makes a payment outside their normal spending category. The mechanism was implemented using HTML, CSS, JavaScript, PHP, and MySQL and could help reduce fraud by adding an additional layer of security validation for online payments. However, it may not detect all fraudulent transactions accurately as the Hidden Markov Model requires at least 10 prior transactions to properly classify users.
2. Contents
• Introduction
• Problem Definition
• Proposed Solution
• Block Diagram
• Implementation
• Software and Hardware Requirements
• Benefits
• Results and Conclusion
3. Introduction
• Online Shopping – one of the largest and
fast going trend
• Mode of payment – credit card, debit card,
Net Banking
• Online payment does not require physical
card
• Major Risk – credit /debit card detail is
known to other
4. Problem Definition
• Online payment does not require physical
card
• Anyone who know the details of card can
make fraud transactions
• Currently, card holder comes to know only
after the fraud transaction is carried out.
• No mechanism to track the fraud
transaction
5. Proposed Solution
• A mechanism is developed to determine
whether the given transaction is fraud or not
• The mechanism uses Hidden Markov Model
to detect fraud transaction
• Hidden Markov Model works on the basis of
spending habit of user.
• Classifies user into Low, Medium or High
category
7. Implementation
• Project is implemented using following technologies :
HTML, CSS, JavaScript, PHP and MySQL
• HTML and CSS is used for interface designing
• JavaScript is used for client side validation
• PHP is used for server side scripting
• MySQL is used for database
8. Hardware & Software Req.
Online
Auction System
• Pentium Core 2
Duo processor or
above
• I GB RAM
• 20 GB HDD
• Router for Internet
Connection
• Windows 2000/
Windows XP/
Windows Vista/
Windows 7
• WAMP
• Macromedia
Dreamweaver
9. Benefits
• Reduction in number of fraud transaction
• User can safely use his credit / debit card for
online transaction
• Added layer of security
10. Results and Conclusion
• Fraud detection is based on Hidden Markov
Model which is learning algorithm, hence not
100% correct
• It has detected those transaction as fraud
where user belongs to low category and high
category payment is made or vice versa
• The mechanism require at least 10 transaction
to determine accurately the transaction as
fraud or not.