Evaluation of performance of machine learning methods in mining structure–property data of halide perovskite materials

Abstract

With the rapid development of artificial intelligence and machine learning (ML) methods, materials science is rapidly entering the era of data-driven materials informatics. ML models serve as the most crucial component, closely bridging material structure and properties. There is a considerable difference in the prediction performance of different ML methods for material systems. Herein, we evaluated models of three categories (linear, kernel, and nonlinear method), with twelve ML algorithms commonly used in the materials field. In addition, halide perovskite was chosen as an example to evaluate the fitting performance of different models. We constructed a total dataset of 540 halide perovskites and 72 features, with formation energy and bandgap as target properties. We found that different categories of ML models show similar trends for different target properties. Among them, the difference between the models is enormous for the formation energy, with the coefficient of determination (R2) range: 0.69~0.953, while the fitting performance between the models whose R2 range: 0.941~0.997 is closer for bandgap. The nonlinear-ensemble model shows the best fitting performance for both the formation energy and the bandgap. It shows that the nonlinear-ensemble model, constructed by combining multiple weak learners, effectively describes the nonlinear relationship between material features and target property. In addition, the eXtreme gradient boosting decision tree model performs the most superior results among all the models and the searching of two new descriptors that are crucial for formation energy and bandgap. Our work provides useful guidance for the selection of effective machine learning methods in the data-mining studies of specific material systems.

Entropy-Driven Stabilization of Multielement Halide Double-Perovskite Alloys

“…, Yb, Ho, Er, etc. ). − The introduction of these elements can provide new potential opportunities to design novel functional materials and devices for optoelectronics, spintronics, thermoelectrics, etc.…”

Machine-learning-assisted discovery of perovskite materials with high dielectric breakdown strength

“…The results of the average 10-fold CV eliminated the randomization and we used the “random_state” with the closest average effect as the final hyperparameter. 38…”

“…The results of the average 10-fold CV eliminated the randomization and we used the ''random_state'' with the closest average effect as the final hyperparameter. 38 After finishing the prediction of phonon cutoff frequency (o), we added the prediction results into the feature set to build the preliminary model for predicting the dielectric breakdown strength of the materials. In order to improve the prediction accuracy and efficiency, we used the traversal method to screen out the best parameters of the model, including the appropriate features, the number of features and the proportion of the training set and test set.…”

Machine‐Learning Accelerating the Development of Perovskite Photovoltaics

“…ML have also been applied to predict properties thermodynamic properties such as atomization energies [ 96 ] and formation energies. [ 97,98 ] Electronic properties, such as carrier effective mass, [ 99 ] conductivity, [ 100 ] and density of states, [ 101 ] have also been the predicted results of ML. Moreover, ML also plays a crucial role in predicting the ion adsorption energy, [ 102 ] hysteresis and reproducibility, [ 103 ] area‐specific resistance, [ 104 ] etc.…”