A Method for Evaluating the Green Economic Efficiency of Resource-Based Cities Based on Neural Network Improved DEA Model

In this study, we use BP neural network to improve the DEA model to conduct in-depth research and analysis on the method of green economic efficiency evaluation of resource-based cities. The traditional DEA cannot make ranking and analysis of effective units, which affects the accuracy of empirical analysis. Accordingly, the BP-DEA model is introduced to further conduct a comparative eco-efficiency analysis of relatively effective provinces. In this study, the optimal inputs and outputs are calculated by DEA, and further, the BP neural network is used to fit the functional relationship between the optimal inputs and outputs, and by adding variables, the trained neural network can be used for the prediction of the optimal outputs. In this study, the BP-DEA model is used to empirically investigate the temporal evolution trend, spatial differences, and efficiency differences in eco-efficiency. Meanwhile, breaking through the limitation that DEA can only calculate regional efficiency values, this study combines the Malmquist index to compare and decompose the eco-efficiency of different provinces to analyze the sources of total factor productivity changes. The results show that the method can clarify the gap between the actual operation of each indicator and the reference point; it can identify how much room for improvement still needs to be made for each indicator, and it can also determine whether each city should be rewarded or penalized and its specific amount. Finally, based on the evaluation of eco-efficiency and the main constraints, corresponding policy recommendations are proposed. Finally, based on the evaluation results of the BP-DEA method, this study analyzes the overall efficiency improvement of cities in the two study areas in three dimensions: urbanization construction, ecology, and economic development put forward seven types of urban efficiency improvement and propose targeted urban development suggestions according to regional characteristics.