NUS-Fintech-Credit-Scoring-XAI

This is a full stack explainable AI application which allows users to make credit score predictions. It is trained on the german credit scoring dataset, and can use a K-Nearest-Neighbour and Random Forest Algorithm to make predictions, based on user inputs.

Apart from simply making the decision on whether it is a good or bad idea to give a loan to a person with the input traits, it is also able to explain why the decision was made, and which factors played major roles in making this prediction.

Setup Instructions

Prerequisites:

1. Have the latest version of Node.js installed on your local machine. You can check the version by typing node --version into your CLI.

2. Have docker 27 installed on your local machine. You can check the version by typing docker --version into your CLI.

How to Setup?

1. Get the source code onto your local machine. You can either do so with git clone https://github.com/Cikguseven/NUS-Fintech-Credit-Scoring-XAI.git or fork the repository and clone it onto your machine.

2. In the backend directory, perform the following actions: 2.1. Ensure that docker is running on your machine, and that you are in the backend directory. 2.2. Run docker build -t credit_scoring_backend:1.0 . in your CLI. This builds a docker image named credit_scoring_backend, tagged as version 1.0. 2.3. Run docker run -p 8000:8000 credit_scoring_backend:1.0 in your CLI. This creates a docker container based on the image and starts the container. Note that -p 8000:8000 is important as it maps the container's port 8000 to the machine's port 8000, allowing external access to the container.

3. In the frontend directory, perform the following actions: 3.1. Add a .env file, with the variable BACKEND_URL="http://127.0.0.1:8000/". 3.2. Run the command npm install. This will install the dependencies needed to run the frontend. 3.3. Run the command npm run dev. This will start the development version of the frontend on your local machine. Note: to use the frontend in a production environment, use next build and next start to build the application for production usage and start a production server, instead of the dev command.

4. Visit localhost:3000/predict and you should find your app running.