Mine-Intel is a machine learning–powered decision support system for predicting roof fall risk in underground mining. It trains and evaluates multiple models on historical data and uses CatBoost for accurate and reliable predictions.

The platform also includes a voice and chat-based assistant for interactive input updates and a SHAP-based analysis dashboard to explain model behavior, enabling safer and more informed mining operations. Features

• Trains and evaluates multiple ML models

 • Linear Regression

 • Random Forest

 • XGBoost

 • CatBoost (final selected model)

• Automatically compares model performance using evaluation metrics

• Uses CatBoost for final prediction due to best accuracy and stability

• Predicts roof fall rate based on mining operational parameters

• Supports real-time input modification before prediction

• Interactive interface for testing new scenarios

• Allows users to update input values via

 • Voice commands

 • Chat interface

• Makes the system usable even for non-technical users in field environments

• Dedicated SHAP analysis page

• Visualizes feature importance and contribution

• Helps engineers understand why a certain prediction was made

• Python • Scikit-learn • CatBoost • XGBoost • SHAP

• Flask • REST APIs

• HTML • CSS • JavaScript

• Matplotlib • Plotly

• SpeechRecognition • Web Speech API

• Git • GitHub • VS Code • Jupyter Notebook

https://github.com/chhavikapasiya11/The-MIne-Intel
cd Techo_Medicine

cd backend
pip install -r requirements.txt
python app.py

cd ../frontend
npm install
npm start

https://github.com/chhavikapasiya11/The-MIne-Intel/blob/main/ss/ss6.png

• Integrate real-time IoT sensor data from mining equipment for live risk monitoring

• Develop a mobile application for on-site engineers and supervisors

• Deploy the system on cloud infrastructure for scalability and remote access

MIT License