Jordan Networks

• 

  Playing around with features

• 

  Signup portal

• 

  Signup portal v2

• 

  Most frauds are conducted in dark hours

• 

  Feature engineering

• 

  Some Interesting facts

• 

  Lost to 20s

• 

  Training model

Inspiration

Elevate AI challenge. We decided to create a solution for them to verify customer before lending money using two machine learning models.

What it does

This program uses two models. The first model tells you if a user is fraudulent based on his or her background and the second model helps prevent fraudulent transactions.

How we built it

We collected datasets on User credibility based on factors like their Cred card History, Employment status, Checking and Saving account status. Three major sources of information are considered: Bank related activities, Credit activities, and Living standards. We also collected datasets on credits loaned to genuine users which failed to pay back either due to genuine reasons or planned deception.

Challenges we ran into

Finding Dataset, Feature engineering, Picking the right model. Training models

Accomplishments that we're proud of

Model accuracy 95%+

What we learned

What's next for Verifying Users to prevent fraud

Built With

• machine-learning
• python