Abstract: Transition metal carbides (TMC) are characterized with high melting points which make experimental determination of their average surface energies a difficult task. A database of 3d, 4d and 5d TMC is hereby established using machine learning technique on the platform of support vector regression (SVR). SVR was built, trained and validated using some selected metals in periodic table and accuracy of 97.5% and 99.2% were achieved during training and testing phase respectively. Average surface energies of TMC were estimated using the trained and tested SVR model. Comparison of our results with surface energies from the first principle calculation and other theoretical results show agreement in terms of absolute values and the trend of variation depicted when Femi energy and density of state are analyzed using linear muffin-tin-orbital (LMTO). The computational ease of this approach in estimating average surface energies of TMC can be an edge over the existing theoretical methods.
 
Key words: Support vector regression, average surface energy, support vector regression model (SVRM) and transition metal carbide.

Cite: Taoreed O. Owolabi, Kabiru O. Akande, and Sunday O. Olatunji, "Estimation of Surface Energies of Transition Metal Carbides Using Machine Learning Approach," International Journal of Materials Science and Engineering, Vol. 3, No. 2, pp. 104-119, June 2015. doi: 10.17706/ijmse.2015.3.2.104-119