Securing financial transactions for banks and providing a safeguard mechanism against fraudulent transactions. Banks currently use a Rule-Based System to red flag possible fraudulent transactions.
What it does
We are developing a Fraud Detection Algorithm using AI to detect and flag Credit card frauds to some extent. Customer’s Needs: Credit Card frauds are very common in Pakistan, to overcome this problem we are developing a solution using AI.
How we built it
We built the Restful API using Machine Learning in Python. The algorithm used for prediction is Logistic Regression
Challenges we ran into
Python Development Environment Issues Thinking largely from a business perspective was difficult because as developers, our mind goes towards the technical side
Accomplishments that we're proud of
We developed a working POC and a small model that would handle the imbalance data sets used in fraud detection and works well on the testing data.
What we learned
We learned the business side of Software Engineering and how to think of your work as a product and not as a project.
What's next for Fraud Detection Using Artificial Intelligence
We would work on improving the Machine Learning model by using a Mesh approach of a combination of classification algorithms and will deploy it on a API and request it to predict incoming data.