Codes

This project is under consideration for a research publication. So till its not publishing, the code and algorithm applied for this code is private". But the basic information about the project is: From the past experiments it can be inferenced that the phase formation of alloys usually depends on the composition of alloy and its thermodynamic parameters. In this project, we employed our own model to predict the phase formation in alloys. We have used the dataset in which three phases are present(AM, SS, IM) in the alloys. The objective is to obtain a relationship between different elemental properties and phases through which the phase of alloy can be predicted by the help of elemental properties. The model is trained to predict the phase of alloy on the basis of input features and achieved an accuracy of 91.66 % with training full dataset. Also, an average accuracy of 79.71 % is achieved in the cross validation training and testing datasets. Model is also compared with other ML classification algorithms on the same dataset and higher accuracy is achieved in comparison to other models.