Abstract
Machine learning (ML) techniques are increasingly recognized as powerful tools for accelerating the discovery of functional materials for energy applications. In this study, multiple supervised ML algorithms, including Random Forest, Support Vector Machine, Naïve Bayes, and Logistic Regression, were implemented to classify and predict oxide-based compounds with potential piezoelectric properties. Among these, the Random Forest model exhibited the best predictive performance, achieving an accuracy of 94.15%, precision of 0.74, recall of 0.99, and a Gini coefficient of 0.93. Guided by ML predictions, BaCO₃ was identified as a promising candidate and further optimized through Ni doping. Experimental validation was carried out via wet-chemical synthesis, and X-ray diffraction (XRD) confirmed the formation of a polycrystalline orthorhombic phase. A clear peak shift to lower angles in Ni-doped samples confirmed successful lattice substitution, while higher Ni concentrations (15%) induced secondary phase formation. This combined computational–experimental approach demonstrates how ML-guided screening significantly reduces cost and time, offering a scalable pathway toward the rational design of energy materials for piezoelectric and dielectric applications.
Article Type
Original Study
First Page
1
Last Page
1
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
G S, Rekha and Sunil, M. Anantha
(2025)
"Studies on Machine Learning-Guided Prediction and Experimental Validation of Ni-Doped BaCO₃ for Energy Applications,"
Eurasian Journal of Physics and Functional Materials: Vol. 9:
No.
4, Article 4.
DOI: https://doi.org/10.69912/2616-8537.1261
