Leveraging Composition-Based Material Descriptors for Machine Learning Optimization

Abstract

In this study, we evaluate several classifiers and focus on selecting a minimal set of appropriate material features. Our objective is to propose and discuss general strategies for reducing the number of descriptors required for material classification. The first strategy involves testing whether the critical temperature of the target material property is invariant with respect to binary groups of composition-based features. We also propose a multi-objective optimization procedure to reduce the set of composition-based material descriptors. The latter procedure is found to be particularly useful when applied to Bayesian classifiers. We test the proposed strategies focusing on low-temperature superconductors material data extracted from a public database.

arXiv comment34 pages, 11 figure panels in the main document, 5 figure panels in the supp. info, preprint

References in corpus (14)

• Scikit-learn: Machine Learning in Python
• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• SMOTE: Synthetic Minority Over-sampling Technique
• A General-Purpose Machine Learning Framework for Predicting Properties of Inorganic Materials
• Machine learning modeling of superconducting critical temperature
• Unconventional Superconductivity
• Deep Learning Model for Finding New Superconductors
• Predicting new superconductors and their critical temperatures using unsupervised machine learning
• Superconducting materials classes: introduction and overview
• Minimal set of crystallographic descriptors for sorption properties in hypothetical Metal Organic Frameworks: Role in sequential learning optimization
• Clustering Superconductors Using Unsupervised Machine Learning
• Quasi Equilibrium Grid Algorithm: geometric construction for model reduction
• Adaptive simplification of complex multiscale systems
• Approximation of slow and fast dynamics in multiscale dynamical systems by the linearized Relaxation Redistribution Method

Cited by in corpus (2)

• Classification-based detection and quantification of cross-domain data bias in materials discovery
• Learning effective good variables from physical data