preetesh21/abs

ABS

Albus Banking System This app was built by me as a practice app for MERN Stack.

For this project I have used MongoDB for storing the user data.Express in the backend and React in the frontend. There is a second server running as well which contains the python model of the Loan Risk Prediction System which classifies if the person would get a loan or not.

An initial attempt was made by me to make the machine learning model using the Tensorflow JS framework but the accuracy of the model was even below a normal guess. So had to eventually give up on that line of work and use Flask and sklearn in python to accomplish the task. Here is animated GIF of the app:

Tech Stack

• Frontend

    React
    Bootstrap
    Sass
    CSS
  

• Backend

    Express JS
    Flask
  

• Database

    MongoDB
  

• Server/Runtime

    Node JS
  

How to run locally

• Install Node and MongoDB(if not using cloud platform) on your system.
• Clone the repository on your system.
• Move into the server folder and Run the command in the terminal

npm install
nodemon app.js

• Then open another terminal and run the commands after entering the directory client

npm install
npm start

• Then open another terminal and run the following commands in flask directory

pip install requirements.txt
python app.py

• Visit "http://localhost:3000" on your browser
• Now you are all set to start!

Database Models Used

There are two models defined for storing data

1.User model

username :{
        type : String,
        required : true,
        min : 6,
        max : 15
    },
    password : {
        type : String,
        required : true
    },
    role : {
        type : String,
        enum : ['user','admin'],
        required: true
    },

2.Bank Form Details Model

gender : {
        type : Number,
        default:0,
        required : true
    },
    marital : {
        type : Number,
        default:0,
        required : true
    },
    dependents : {
        type : Number,
        default:0,
        required : true
    },
    education : {
        type : Number,
        default:0,
        required : true
    },
    status : {
        type : Number,
        default:0,
        required : true
    },
    appincome : {
        type : Number,
        default:0,
        required : true
    },
    coapincome : {
        type : Number,
        default:0,
        required : true
    },
    loan : {
        type : Number,
        default:0,
        required : true
    },
    loanterm : {
        type : Number,
        default:0,
        required : true
    },
    credit : {
        type : Number,
        default:0,
        required : true
    },
    property : {
        type : Number,
        default:0,
        required : true
    },
    predict:{
        type:Number,
        default:0,
        required: true
    }

Machine Learning Model

For the machine learning part several models were tried including Decision Tree,SVM and K-Means.The best result came through Logistic Regression.The data used here is taken from kaggle and comprises of only two categories with the distribution as 65% and 35% which makes the minimum accuracy to be 65% for a model to be considered for evaluation. The data composed of deatils such as:

['Gender','Married','Dependents','Education','Self_Employed','Credit_History','Property_Area','Loan_Status']

Sklearn package was used for machine learning. Data cleaning was done prior to fitting the data to the model.

Model	Accuracy	Time
Logisitic Regression	80.803%	0.7364733219146729 s
GNB	79.803%	0.015659332275390625 s
Random Forest	74.384%	0.06247997283935547 s
Decision Tree	75.369%	0.015588045120239258 s
SVC	65.517%	47.450788736343384 s
KNN	60.099%	0.1722092628479004 s
MLP	53.695%	0.48421430587768555 s
AdaBoost	79.310%	0.1476762294769287 s

Workflow

So the app workflow is as follows: First the user needs to register himself(either as an admin or a regular user) and then using those credentials he needs to log in which would lead him to the main page which contains the form details for the loan. Now after filling the details he would be informed whether his credentials land him a loan or not. If the user is an admin then he has an additional capability to view all the form details filled uptill now.

Directory Structre

|-- README.md
|-- client
|   |-- node_modules
|   |-- README.md
|   |-- package-lock.json
|   |-- package.json
|   |-- public
|   |   |-- background.jpg
|   |   |-- favicon.ico
|   |   |-- index.html
|   |   |-- logo192.png
|   |   |-- logo512.png
|   |   |-- manifest.json
|   |   `-- robots.txt
|   |-- src
|   |   |-- App.js
|   |   |-- Components
|   |   |   |-- Admin.js
|   |   |   |-- Home.js
|   |   |   |-- Loans.js
|   |   |   |-- Login.js
|   |   |   |-- Message.js
|   |   |   |-- Navbar.js
|   |   |   |-- Register.js
|   |   |   `-- Unauthorized.js
|   |   |-- Context
|   |   |   `-- AuthContext.js
|   |   |-- Services
|   |   |   |-- AuthService.js
|   |   |   `-- LoanService.js
|   |   |-- hocs
|   |   |   |-- PrivateRoute.js
|   |   |   `-- UnPrivateRoute.js
|   |   |-- index.js
|   |   |-- react-app-env.d.ts
|   |   `-- unauthorized.scss
|   `-- tsconfig.json
|-- flask
|   |-- __pycache__
|   |   `-- app.cpython-38.pyc
|   |-- app.py
|   |-- classifier.py
|   |-- data.csv
|   |-- model
|   |   |-- AdaBoost.pkl
|   |   |-- Decision_Tree_Classifier.pkl
|   |   |-- Gaussian_Naive_Bayes.pkl
|   |   |-- KNN.pkl
|   |   |-- Logistic_Regression.pkl
|   |   |-- Multi_Layer_Perceptron.pkl
|   |   |-- Random_Forest.pkl
|   |   `-- SVC_Linear.pkl
|   |-- requirements.txt
|   `-- templates
|       `-- index.html
|-- gif.gif
`-- server
    |-- node_modules
    |-- app.js
    |-- models
    |   |-- Loan.js
    |   `-- User.js
    |-- package-lock.json
    |-- package.json
    |-- passport.js
    `-- routes
        `-- User.js