Resampling Techniques for Materials Informatics: Limitations in Crystal Point Groups Classification.

Imbalanced data sets in materials informatics are pervasive and pose a challenge to the development of classification models. This work investigates crystal point group prediction as an example of an imbalanced classification problem in materials informatics. Multiple resampling and classification techniques were considered. The findings suggest that the most influential variable of the resampling algorithms is the one controlling the number of samples to omit (undersample) or synthetically generate (oversample), as expected. The effect of balancing is to enhance the classification performance of the minority class at the cost of reducing the correct predictions of the majority class. Moreover, ideal balancing, where the classes are precisely balanced, is not optimum. Alternatively, partial balancing should be performed. In this study, the ideal ratio of the minority to majority class was found to be around two-thirds. The biggest improvement in the classification was for the random undersampling technique with k-nearest neighbors and random forest.

Highly accurate machine learning prediction of crystal point groups for ternary materials from chemical formula.

One of the most challenging problems in condensed matter physics is to predict crystal structure just from the chemical formula of the material. In this work, we present a robust machine learning (ML) predictor for the crystal point group of ternary materials (A[Formula: see text]B[Formula: see text]C[Formula: see text]) - as first step to predict the structure - with very small set of ionic and positional fundamental features. From ML perspective, the problem is strenuous due to multi-labelity, multi-class, and data imbalance. The resulted prediction is very reliable as high balanced accuracies are obtained by different ML methods. Many similarity-based approaches resulted in a balanced accuracy above 95% indicating that the physics is well captured by the reduced set of features; namely, stoichiometry, ionic radii, ionization energies, and oxidation states for each of the three elements in the ternary compound. The accuracy is not limited by the approach; but rather by the limited data points and we should expect higher accuracy prediction by having more reliable data.