Identification method of nonlinear dynamic system based on neural network

The invention provides an identification method of a nonlinear dynamic system based on a neural network. The identification method is divided into three function levels of a data bottom layer, data analysis and data display through a BI system. The data bottom layer is responsible for managing data, and the data bottom layer comprises data acquisition, data ETL, data warehouse construction, data display, report presentation and visual chart data solution for data display, and provides personalized precise commodity pushing according to user purchase feature behavior data; on the other hand, a marketing promotion channel is optimized, exposure and passenger flow are obtained efficiently at low cost, and sales optimization is carried out in combination with an intelligent real-time pricing strategy; according to the method, the mathematical model of the object is estimated by measuring the input and output states of the object, so that the established mathematical model and the controlled object have the same i