Search

With the development of information technology and the deepening of government system reform, e-government has become one of the key points of government construction. However, both in developing countries and developed countries, there is a common phenomenon that the success rate of e-government projects is not high. Although various e-government evaluation index systems have been proposed in both theoretical research and application aspects, which provide strong support for the development of e-government, these index systems still have limitations. This paper comprehensively applies BP neural network and adaptive matching tracking, and constructs an e-government performance evaluation model based on adaptive matching neural network. We conduct simulation experiments on the constructed model to verify the accuracy and rationality of the model evaluation. It is feasible to introduce the Balanced Scorecard to construct the E-government performance evaluation index system. The constructed index system has strong objectivity, operability, comprehensiveness, guidance and sustainability, and can be applied to the practice of e-government performance evaluation. It is feasible to use the adaptive matching tracking neural network model to evaluate the performance of e-government. The model has the ability to self-learn the evaluation samples, and can grasp the mapping relationship between the evaluation indicators and the evaluation results, so as to imitate the evaluation of experts, and overcome the evaluation randomness and subjective uncertainty that are difficult to get rid of in traditional manual evaluation. In order to better evaluate the operation status of the e-government system and its adaptability to future needs, and find out its existing problems, this paper fully considers the interaction process between the government and enterprises, and establishes a corresponding responsive government. The application of adaptive matching tracking neural network model for e-government performance evaluation has better optimization results than the original BP neural network model. The specific performance is that the convergence speed is fast, the training will not fall into a local minimum, and the error of the evaluation results is small. [ABSTRACT FROM AUTHOR]