Method for quickly predicting specific surface area of perovskite material based on XGBoost algorithm

The invention relates to a method for quickly predicting the specific surface area of a perovskite material based on an XGBoost machine learning algorithm, and the method comprises the steps: searching the specific surface area data and chemical formula of an ABO3 type perovskite material from literature through investigating the literature, and enabling a preprocessed data set sample to serve asa data set sample; generating a characteristic variable by utilizing the sample set; randomly dividing the data set sample into a training set and a test set; the specific surface area of an ABO3 typeperovskite material sample is taken as a target variable, and part of obtained characteristic variables are taken as independent variables. Establishing a rapid forecasting model of the specific surface area of the ABO3 type perovskite material by adopting an XGBoost algorithm; screening the characteristic variables, and establishing a forecasting model; and forecasting the specific surface areaof the obtained test set sa