Skip to content

lihomm/FYP

Repository files navigation

Payment Transaction Fraud Detection System

Overview

This is my Final Year Project submitted to Asia Pacific University as part of the fulfillment of my Bachelor's Degree in Computer Science. There were 4 machine learning models developed in this project namely, Deep Neural Network, Long Short-Term Memory (LSTM), Random Forest, LightGBM to predict frauds in payment transactions. The payment transaction data was supplemented by Vesta Corporation at here.

Key Features

  1. Model Agnostic Explainability
    The model's prediction can be explained using the LIME value in the dashboard, hence this project is under the category of Explainable AI (XAI) for responsible and ethical AI use in the commercial landscape.

  2. Advanced Data Engineering
    Advanced techniques were used to handle the dataset that presented great challenge to achieve respectable score for the model in terms of AUC-ROC and AUC-PRC.

  3. Intuitive Dashboard Deployment
    The prediction made by the model can be viewed all at once on the dashboard deployed using Streamlit as frontend and FastAPI as the backend.

Installation

To get started with the project, follow these steps:

  1. Clone the repository: git clone https://github.com/lihomm/FYP.git

  2. Navigate into the project directory: cd FYP

  3. Navigate into the backend directory: cd backend

  4. Install dependencies listed in the requirements.txt file: pip install -r requirements.txt

  5. You are good to go!