The use of principal component analysis (PCA) and partial least square (PLS) for designing new hard inverse perovskites materials

Abstract

The present article is an investigation on the physical and mechanical properties of cubic inverse carbides and nitrides perovskites by means of data mining; an excellent method for examining and discovering hidden knowledge within data. The goal of the data mining technique is to detect and classify the consequences of existing property relationships and discover the implicit, yet meaningful, relationship between the data set elements locating the correlations between the properties of materials. Multivariate modeling methods, such as principal component analysis (PCA) and partial least squares regression (PLS) are used to present the investigation findings. In this study, we refer to multivariate PCA and PLS methods to predict a robust model of a lattice constant in a cubic inverse carbide and nitride perovskite having the formula ABX3 based on the atomic radius of the different species. Data mining is an ideal approach to extract the most possible information from the available data. Also, we present a data mining approach in order to predict new ultra hard coatings materials and another approach to predict which compounds among inverse perovskites possess the potential to achieve high hardness and fracture toughness for applications as a thermal barrier coating. These results may serve as a map for the design of inverse perovskite-related new multilayer ultra-hard coating materials.