🏦 Loan Approval Predictor Bot 🤖

🚀 Overview

This Streamlit-based web application helps users predict whether their loan application is likely to be approved or rejected based on their financial and demographic information. Built as a Capstone Project, this bot uses a pre-trained machine learning model (Logistic Regression with SMOTE for class balancing) and natural language processing to extract insights from user inputs and provide meaningful predictions.

The app also uses OpenAI's LLM to explain the model’s decision in plain English, making it both intelligent and user-friendly.

---

✨ Features

• 🔍 Loan Approval Prediction using a pre-trained ML model.
• 🧠 NLP-based Input Parsing: Enter financial info in natural language, and the app extracts key features automatically.
• 📊 Loan Eligibility Calculator: Get insights into how much loan you can afford based on your income and EMI.
• 📈 Dynamic Visuals: Interactive UI built with Streamlit for a smooth user experience.
• 🤖 Explainable AI: Model decisions explained using OpenAI’s language model.
• 📦 Categorical Feature Encoding: All categorical inputs are automatically label-encoded to match model expectations.
• 🏗️ Box-Cox Transformation: Applied on income feature for better model performance.

---

🧠 Tech Stack

• Frontend & App: Streamlit
• Machine Learning: Scikit-learn (Logistic Regression with SMOTE)
• NLP: spaCy (en_core_web_sm)
• LLM Integration: OpenAI GPT via langchain
• Model Deployment: GitHub-hosted .joblib model file
• Libraries: pandas, numpy, joblib, requests, spacy, streamlit, openai, langchain

---

📂 Directory Structure

Loan-Approval-Predictor/
├── Loan_app.py # Main Streamlit app script
├── Capstone_PreProcess_DKP_Final.ipynb # Preprocessed data
├── Capstone_ML_Modelling_DKP-Final-Copy2.ipynb # ML model training
├── Logistic2_Smote_Model.joblib # Pre-Trained model (auto-downloaded)
├── README.md # Project overview
└── requirements.txt # Python dependencies

---

🔧 Setup Instructions

1. Clone the repository

git clone https://github.com/dkamp007/Capstone.git
cd Capstone

2. Install Dependencies

pip install -r requirements.txt
python -m spacy download en_core_web_sm

3. Set your OpenAI API Key

Create a .streamlit/secrets.toml file:

apikey = "your_openai_api_key"

4. Run the app

streamlit run Loan_app.py

Check out the webapp!