Dagger

Inspiration

We noticed that the banking industry is one of the most prominent industries where many frauds occur. Therefore, we wanted to address a problem that affected many people and develop a solution to mitigate the inconveniences of financial purchases/fraud.

What it does

Our product Dagger helps banks input customer transactional data and our machine learning algorithm successfully identifies if and where a financial fraud occurred. Using normal distribution, global trend, and expectation values, we created a model that was able to closely minick and detect the normal causes of fraud, like high volume purchases, time of day, and the distance from the user's home.

How we built it

To build Dagger we had two components the front end and the back end. To build the backend we used Python to train our ML model and we used a linear regression algorithm and trained it on custom synthetic data. Furthermore, we used Python libraries to create charts that visually represent our results which showcase the predicted financial fraud. The front end used HTML, CSS, and a local flask, and was able to constantly update real-time data, based on specific user transactions.

Challenges we ran into

Some challenges we ran into were that initially when building our machine learning model we had to teach the program how to differentiate between fraud and nonfraud. However, this was difficult because when we were analyzing our model we realized that there was almost no difference between fraud and legit data. This setback caused us to re-train our model until the results matched global banking trends.

Accomplishments that we're proud of

We're proud of creating a machine learning model that could potentially have a significant impact on the financial industry, especially in the growing world of connectivity and the rise of security concerns. We're also proud to have networked with many recruiters and were amazed by the unique challenges they posed, especially PNC bank and Mr.Muthu helped us approach the problem in an innovative way we previously didn't even think of.

What we learned

We learned about the importance of the integration of front and back ends, especially in order to make a product that's large-scale and multifunctional. In particular, we all learned how machine learning and generative AI are prominent and expanding uses of security and reliance in the finance and banking industry.

What's next for Dagger

The next thing for Dagger is to have generative AI feedback regarding the frauds identified. It gives the user real-time feedback on their transactions and explains why a fraud detection has occurred.