Iterative Stacking Regression for Materials Properties Prediction

Most common data-driven approach for material analysis are the regression models that estimate specific material properties as a function of other related physical quantities. Data for the estimated targets and the input variables are assumed to be available.In this paper we go beyond the baseline regression approach and propose an iterative stacking regression model to subsequently estimate material properties that are difficult to measure and therefore not available for all data samples. The proposed approach is inspired by the stacking ensemble learning that combines the predictions of multiple ML models to improve the regressor performance. Overall this work provides valuable alternative how to deal with regression problems when data is small and incomplete.