ML-assisted-alloy-design-using-wrought-aluminum-alloys

KO | EN

gitlite — search

#golang #android #rust #linux #docker #macos #windows #awesome #cli #kubernetes #awesome-list #vue #pytorch #mcp

ML-assisted-alloy-design-using-wrought-aluminum-alloys

(★ 10)

codes and dataset used in the study "A feasibility study of machine learning-assisted alloy design using wrought aluminum alloys as an example" by Yasaman J. Soofi, Md Asad Rahman, Yijia Gu, Jinling Liu

ML-assisted-alloy-design-using-wrought-aluminum-alloys Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# ML-Assisted Alloy Design Using Wrought Aluminum Alloys

This repository contains the code and dataset used in the study:  
**"A feasibility study of machine learning-assisted alloy design using wrought aluminum alloys as an example"**  
by **Yasaman J. Soofi, Md Asad Rahman, Yijia Gu, and Jinling Liu.**

The study explores the application of **machine learning** to predict and analyze material properties, offering a data-driven approach to alloy design. It highlights the use of **advanced regression and classification models** to understand relationships between composition, processing, and resulting mechanical and technological properties.

[Read the published paper here.](https://www.sciencedirect.com/science/article/pii/S0927025622004943)

---

## Overview

This repository features Python code to perform **predictive modeling**, **feature analysis**, and **validation** of machine learning models, using datasets focused on wrought aluminum alloys. The project demonstrates the potential of ML techniques like **Random Forest, XGBoost, SVR**, and **ANN**, combined with advanced feature engineering and statistical validation, to enhance alloy design.

Key highlights include:
- Achieved an **8% improvement in prediction accuracy** (R² from 0.9 to 0.97) through hyperparameter tuning and advanced feature engineering.
- Expanded dataset size and improved model robustness using web scraping techniques (**BeautifulSoup, Selenium**).
- Validated model performance using **cross-validation**, **random seed testing**, and statistical methods (**t-tests**), with actionable insights from **feature importance analysis**.

---

## Code Files

### Statistical Analysis
1. **`correlation_heatmap.ipynb`**  
   - Performs Spearman correlation analysis and generates a heatmap to visualize relationships between variables.

2. **`feature_importance.ipynb`**  
   - Analyzes feature importance for mechanical and technological properties. Produces figures showcasing the top-10 important features for each property.

### Predictive Modeling
3. **`100seeds_mechanical.ipynb`**  
   - Implements regression models (e.g., Random Forest, XGBoost, SVR) to predict mechanical properties, trained across 100 random seeds for robustness.

4. **`100seeds_technological.ipynb`**  
   - Implements classification models (e.g., Random Forest, ANN) to predict technological properties, trained across 100 random seeds.

### Validation and Visualization
5. **`LOOCV_scatterplt_mechanical.ipynb`**  
   - Trains and evaluates regression models using Leave-One-Out Cross-Validation (LOOCV) and generates scatter plots for mechanical property predictions.

6. **`LOOCV_technological.ipynb`**  
   - Trains and evaluates classification models using LOOCV for technological properties.

---

## Key Features
- **Supervised Learning**: Random Forest, XGBoost, SVR, ANN.
- **Web Scraping**: BeautifulSoup and Selenium for data collection from online repositories.
- **Data Preprocessing**: Ensured high data quality using Python libraries such as pandas and numpy.
- **Visualization**: Correlation heatmaps, feature importance plots, and scatter plots created using Matplotlib.

---

## Technologies Used
- **Programming Languages**: Python (pandas, numpy, scikit-learn, Matplotlib, TensorFlow).  
- **Statistical Methods**: Cross-validation, LOOCV, t-tests, and random seed testing.  
- **Machine Learning Frameworks**: scikit-learn, TensorFlow.  
- **Web Scraping**: BeautifulSoup, Selenium.

---

## Related Work

For further advancements in machine learning for alloy design, explore my subsequent project:  
[An Adaptive Physics-Based Feature Engineering Approach for Machine Learning-Assisted Alloy Design](https://github.com/yasamanjs/An-Adaptive-Physics-based-Feature-Engineering-Approach-for-Machine-Learning-assisted-Alloy-Disc).

---

## Citation

If you use this repository in your work, please cite:  
**"A feasibility study of machine learning-assisted alloy design using wrought aluminum alloys as an example"**  
by **Yasaman J. Soofi, Md Asad Rahman, Yijia Gu, and Jinling Liu**.  
[DOI Link](https://www.sciencedirect.com/science/article/pii/S0927025622004943)

---

ML-assisted-alloy-design-using-wrought-aluminum-alloys

ML-assisted-alloy-design-using-wrought-aluminum-alloys Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

ML-assisted-alloy-design-using-wrought-aluminum-alloys

ML-assisted-alloy-design-using-wrought-aluminum-alloys Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README