Prediction And Analysis Of Material Band Gap Based On Regression Model

Perovskite(Ca Ti O3)type compound is a class of functional materials that can be used to produce solar cells,sensors,solid resistors,etc.This paper uses the perovskite oxide related data set in the open source material engineering database Material Project to analyze the Pearson correlation coefficient to find out the ion radius,electronegativity,tolerance factor and octahedron factor,etc.,which affect the perovskite oxide An important factor in the material band gap.The effective atomic radius,first ionization energy,octahedral factor,tolerance factor and other characteristics of A and B ions are selected to improve the accuracy of band gap prediction.The original data set was randomly divided into training set and test set,and three regression models(Lasso regression,support vector regression SVR,XGBoost)were used to predict the perovskite oxide band gap,and the prediction accuracy of the regression model was compared,and The fitting prediction is analyzed for the effect of different model errors.Among them,SVR and XGBoost have better prediction effects.By exploring and constructing a reasonable regression model to improve the prediction accuracy of the band gap,the effectiveness of the relevant regression model for the prediction of the band gap of perovskite oxide is discussed.Exploring how to further improve the efficiency of material prediction and analysis has certain applied research significance.