AI-Powered Loan Demand Matching and Approval Prediction Application Tool

🔍 Matches cases with similar categories and sentiments to the user's loan application through the ZHIPU AI large model. Analyzes Income, Credit Score, Loan Amount, DTI_Ratio, and Employment Status by building a logistic regression model to predict loan approval results for users.

Core Features

• Semantic Search: Input loan application description, category, sentiment → Real-time matching of historical similar demands based on the ZHIPU AI large model case vector database.
• Intelligent Prediction: Comprehensive analysis of key features such as Income, Credit Score, Loan Amount, DTI_Ratio, and Employment Status. Utilizes a logistic regression model trained on cases for prediction, with a theoretical accuracy of over 90%.

Installation Instructions

Clone the project repository:

git clone https://github.com/JP3000/Loan-Or-Not.git
cd .venv/

Install dependencies:

pip install -r requirements.txt

Usage Instructions

Dataset Source: Kaggle

Configure the .env file:

ZHIPUAI_API_KEY=your_api_key

Run the main program:

cd .venv/
streamlit run app.py

Technology Stack

• Application Text Matching: langchain, ZhipuAI, chroma vector database
• Financial Feature Prediction: sklearn logistic regression model
• Demo Presentation: streamlit

Project Demo