Multivariate analysis and classification of bulk metallic glasses using principal component analysis

A high dimensional data set comprising eleven chosen attributes of 594 bulk metallic glasses (BMGs) compositions has been analysed and classified. Each attribute corresponding to a given composition has been expressed in terms of the equivalent parameter, i.e., atomic fraction weighted sum. Principal Component Analysis (PCA) has been employed to perform the multivariate analysis, aiming at the dimensionality reduction of the high dimensional data set. The evolved principal components are evaluated on the basis of the characteristic temperatures of bulk metallic glasses. The analysis of 594 BMG alloys resulted into a classification model of the alloys which exhibits reliable agreement with Inoue's classification developed on the basis of classical subject knowledge. The occasional cases of the BMGs remaining unclassified according to Inoue's scheme have been judiciously designated as the new classes. (C) 2015 Elsevier B.V. All rights reserved.