Material Properties Predictions Using Data-Driven Technology

Artificial intelligence and smart computing are just in the inception phase. With the availability of the computational and experimental database, the unification of analysis, prediction, and discovery has evolved into a prominent focus in alloy exploration. In this using deep learning concepts, a correlation is developed between mechanical properties and alloy composition to achieve significant prediction accuracy of a dataset of low alloy steel taken from the online repository. Backpropagation with the early stopping concept is used to minimize overfitting and predict results with an r2_score of 0.98.