Related Experiment Video
Updated: Jul 12, 2026
10:55
Extracting the Young's Modulus of Native Murine Pulmonary Basement Membranes from Atomic Force Microscopy Derived Force Maps
Published on: January 31, 2025
Predicting the Young's Modulus of biomedical titanium alloys using machine learning: a data-driven approach
Muhammad Shahmir Saif1, Muhammad Ali Siddiqui1, Fahim Raees2
1Department of Metallurgical Engineering, Computational and Experimental Materials Innovation Group (CEMIG), NED University of Engineering and Technology, Karachi, 75270, Pakistan.
Scientific Reports
|July 9, 2026
Summary
This study introduces a machine learning model to predict Young's Modulus (YM) in titanium alloys for medical implants. The Deep Neural Network (DNN) accurately forecasts material properties, aiding in reducing implant stress shielding.
Area of Science:
- Materials Science
- Biomedical Engineering
- Computational Materials Science
Background:
- Biomedical titanium alloys are crucial for implants due to their biocompatibility and mechanical properties.
- Stress shielding, a consequence of stiffness mismatch between implants and bone, can lead to implant failure.
- Accurate prediction of Young's Modulus (YM) is essential for designing titanium alloys that mitigate stress shielding.
Purpose of the Study:
- To develop and validate a machine learning framework for predicting the Young's Modulus (YM) of biomedical titanium alloys.
- To address the challenge of stress shielding in orthopedic and dental implant applications.
- To establish a predictive model capable of handling complex non-linear relationships within Ti-Alloys data.
Main Methods:
- A Deep Neural Network (DNN) model was constructed utilizing nineteen compositional features and physically meaningful descriptors.
- Data preprocessing involved Box-Cox and Yeo-Johnson transformations to optimize data distribution.
- Model training incorporated L1/L2 regularization, dropout, and early stopping to ensure generalization and prevent overfitting.
Main Results:
- The optimized DNN model achieved a Mean Squared Error (MSE) of 0.294 GPa and an r-squared score of 0.82 on the testing dataset.
- Comparative analysis demonstrated superior performance of the DNN model over XGBoost, Random Forest, Gradient Boosting, and Support Vector Machine.
- The model effectively captured the complex, non-linear relationships between alloy composition and Young's Modulus.
Conclusions:
- The developed DNN framework provides a robust and accurate method for predicting the Young's Modulus of biomedical titanium alloys.
- This predictive capability can guide the development of novel titanium alloys with optimized mechanical properties for enhanced implant performance.
- The study highlights the potential of machine learning in accelerating materials discovery and design for biomedical applications.
