Data-Driven Boundary Regression for Equivalent Identification of External Parameters in Steady-State Models of Power Systems

•A decoupled equation with correction terms is established, and the linear expression of the EE network is derived.•An MLR model based on historical boundary information is used to reflect the operating state of the internal network.•The BCT based on MLE is used to solve the normality performance problems of the regression response variable.•A data-driven dynamic equivalent framework is proposed.