Regional carbon emission prediction method based on XGBoost

• Main Authors: WANG GUODONG, ZHANG DI, JIN JIXIN, WANG NING, ZHOU XIAOLEI, LI YANG, BAI XUE, QI BOLIN
• Format: Patent
• Language: Chinese; English
• Published: 06.02.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention relates to a carbon emission prediction method based on XGBoost, and the method comprises the steps: carrying out the preprocessing of night light data, temperature data, administrative boundary vector data and energy emission data, and carrying out the correction processing of the brightness values (DN) of two kinds of discontinuous night light data in time; and inputting the corrected night light data into a PSO-BP neural network model, obtaining a long-time-series night light data set through the PSO-BP neural network model, fusing auxiliary data such as temperature, latitude and longitude and the like to serve as the input of an XGBoost model, and finally performing accounting and prediction of regional carbon emission by using the trained model. According to the method, the advantages of the BP neural network and the XGBoost are fully fused, the accuracy and generalization ability of the model are improved, carbon emission accounting and prediction on long-time sequences of different region