🏗️ Compressive Strength & 🌱 CO₂ Emission Prediction Smart Material Property Prediction using Machine Learning

📌 Overview In this project, we predict the compressive strength of materials and estimate the CO₂ emissions during production using CatBoost with Optuna-based hyperparameter optimization.

Why?

Material engineers need accurate strength predictions to ensure safety and efficiency.

Environmental researchers require CO₂ emission estimation for sustainable production.

Our model aims to help manufacturers, researchers, and decision-makers make data-driven and eco-conscious choices.

🚀 Features

✅ Predicts Compressive Strength of materials

✅ Estimates CO₂ emissions during production

✅ Uses CatBoost for superior accuracy

✅ Optuna for hyperparameter optimization

✅ Clean & reproducible code

✅ Easy-to-use for both research and industry applications

🧠 Methodology

1️⃣ Dataset Preprocessing

- Missing value handling

- Outlier detection & removal

- Feature scaling & encoding

2️⃣ Model Selection

- Chose CatBoost for its handling of categorical data, robustness, and interpretability

3️⃣ Model Training

- Separate models for Compressive Strength and CO₂ Emission prediction

- Train-test split & cross-validation

4️⃣ Model Optimization (Optuna)

- Hyperparameters like depth, learning_rate, iterations tuned for best performance

- Objective function maximizes model accuracy while minimizing RMSE

📊 Results

Compressive Strength Model → High R² score (>0.9)

CO₂ Emission Model → Accurate estimation within low error margins

Optimized models outperform default CatBoost settings

📂 Project Structure

📦 compressive-strength-co2-prediction

┣ 📜 data/ # Dataset files

┣ 📜 notebooks/ # Jupyter notebooks for analysis

┣ 📜 src/ # Main source code

┃ ┣ model_training.py # Model training logic

┃ ┣ optimize_model.py # Optuna optimization

┣ 📜 requirements.txt # Dependencies

┣ 📜 README.md # Project README

🛠️ Tech Stack

Python 🐍

CatBoost 📈

Optuna ⚙️

Pandas, NumPy, Matplotlib, Seaborn

📦 Installation & Usage

Clone the repository

git clone https://github.com/yourusername/compressive-strength-co2-prediction.git

Run training

comparison-4 models.ipynb

Run optimization

CatBoost optimization.ipynb

📌 Future Improvements

🔹 Add SHAP explainability for feature importance

🔹 Include real-time prediction API

🔹 Expand dataset for global applicability

🤝 Contributing

Pull requests are welcome! If you have ideas to make this better, please fork the repo and submit a PR.