blog.1.image

Fraud Detection in Online Transactions

The project aimed at detecting fraud online transactions based upon transaction parameters such amount, bill number, time of transaction, etc.

  • Domain
    Artificial Intelligence
  • Industry
    Fintech
  • Year
    2019
  • Tech Stack
    Deep Learning
    Django

Project Overview

The project was to identify the fraud online transactions based on various parameters like the time of the transaction, location, bill number, amount, amount cardholder billing, etc. The normal approach would be to record and process previous customers' transactions' details in order to detect fraud transactions. But, the out of the box measure of dealing with such transactions would be to process the details entered by the payor in real-time and give an immediate output of whether it is a valid transaction or a fraud transaction.

 

Need for the Project

We live in an era of advanced technology, where we want everything to give us results by just pressing buttons instantly. There are numerous online platforms from where we can purchase products starting from our daily needs to glamourous things. For that, we have online payment gateways where we can confirm the process of shopping.

As consumers take advantage of speed, ease, and convivence of managing their finances, the fraudsters have also evolved their methods of attack. It is necessary to create the right balance on providing a delightful customer experience while protecting and safeguarding them. This results in the concept of fraud detection in online transactions.

 

Benefits

  • The use of transactional parameters rather than customer profiles
  • Safeguarding customer privacy
  • Immediate alert for fraud transactions

 

Implementation Method

The implementation is done in the following steps:

  1. Data preprocessing
  2. Applying deep learning models
  3. Integrating with Django web app

Outcomes

We successfully developed a Django based web application that distinguishes between fraud and not fraud online transactions based on the transaction data provided by the customer.

 

The accuracy obtained for the model to work was 96%.

Shelley McNabb avatar image
We were able to kick start our business thanks to the optimum and timely delivery by team Ohuru.”
Shubankar
Salozone
Mary Goddard avatar image
“Brilliant team, delivers great solutions with cutting edge technologies."
Sanjeev Duggal
Retd. GM, SAIL
Bradley Singleton avatar image
“True professional firm, proved to be a very valuable asset to our organization.”
Vikhyat Chauhan
The Next Move

Get our services today

Please contact us for more details on these projects or to avail our services.