This project focuses on predicting the percentage of silica concentrate in a mining flotation process using Machine Learning. By analyzing sensor data collected from the flotation plant, the model helps estimate product quality before laboratory testing, enabling faster decision-making and improved process efficiency.
This project was completed as Project 10 of the UCT Machine Learning Internship.
- Predict % Silica Concentrate
- Improve mining process quality
- Reduce impurities in ore concentrate
- Support predictive manufacturing
- Assist engineers in making data-driven decisions
Dataset: Mining Process Flotation Plant Database
Source: UCT Internship Dataset
- % Iron Feed
- % Silica Feed
- Starch Flow
- Amina Flow
- Ore Pulp Flow
- Ore Pulp Density
- Ore Pulp pH
- Air Flow Sensors
- Level Sensors
- % Iron Concentrate
Target
- % Silica Concentrate
- Python
- Pandas
- NumPy
- Matplotlib
- Seaborn
- Scikit-learn
- Joblib
- HistGradientBoostingRegressor
| Metric | Score |
|---|---|
| MAE | 0.2799 |
| MSE | 0.1450 |
| RMSE | 0.3807 |
| R² Score | 0.8853 |
- Correlation Heatmap
- Actual vs Predicted
- Residual Plot
- Target Distribution
Quality-Prediction-in-a-Mining-Process/
│
├── graphs/
├── dataset/
├── Quality_Prediction_in_a_Mining_Process.ipynb
├── mining_quality_prediction_model.pkl
├── requirements.txt
└── README.md
The HistGradientBoostingRegressor model achieved an R² Score of 0.8853, demonstrating strong predictive performance on industrial mining data.
Completed as part of the UCT Machine Learning Internship.
Jayachandiran K
B.Tech Artificial Intelligence & Data Science
Nehru Institute of Engineering and Technology