Fraud-Detection-with-ML

Publish Site: https://xzhangfox.github.io/Fraud-Detection-with-ML/ This page for IT consulting company website simulation. It aims to present two data cases from the first perspective of FoxData.

This project is split into two cases:

• The first case aims to implement supervise model to detect fraud transactions.
• The second case focus on data processing and visualizations.

Both cases are visualized by D3, Plotly, and Matplotlib.

Nootbook

For details of data processing and methodologies. Please visit the shared Google Colab Notebooks:

Case 1: Fraud Detection

Case 2: Customer Data Processing

Dependencies

• Python 2.7/3.7
• Visual Stiduo Code
• imblearn 0.5.0+
• sklearn
• D3

Motivation

The project is an exercise in challenging full-stack positions. Aimed at completing a series of comprehensive qualities including data preprocessing, modeling, visualization, front-end design and so on in a short time.Tried multiple D3 diagrams at the same time, understood the architecture, and skillfully implemented. The total time was 15 hours.

Future Improvement

• Try more machine learning models and manually tune to get the optimal solution.
• Optimize the problem of too long single page length for better user interaction.
• Embed model allows the user to directly play with new inputs. Make the webpage have the function of real-time calculation and feedback.