基于EFA-GA-XGBoost组合预测模型的绝缘子表面污秽程度预测方法  

Prediction method of insulator surface pollution degree based on EFA-GA-XGBoost combined prediction model

作　　者：赵昕迪 Zhao Xindi(Chongqing University of science and technology,Chongqing,401331)

机构地区：[1]重庆科技学院,重庆401331

出　　处：《电子测试》2022年第5期68-70,共3页Electronic Test

摘　　要：文章提出了基于探索性因子分析的GA-XGBoost组合预测模型评估绝缘子表面污秽状态,以个因子变量及其得分作为输入,绝缘子上下表面的ESDD和NSDD作为输出建立基于XGBoost的污秽预测,所提模型具有更短的预测时间、更高的预测精度和更强的泛化能力。A GA-XGBoost combined prediction model based on exploratory factor analysis is proposed to evaluate the pollution state of insulator surface.The pollution prediction based on xgboost is established with factor variables and their scores as input and ESDD and NSDD on insulator upper and lower surfaces as output.The proposed model has shorter prediction time,higher prediction accuracy and stronger generalization ability.

关 键 词：EFA-GA-XGBoost 预测模型 

分 类 号：TM615[电气工程—电力系统及自动化]